Wind farm reactive power coordination method and device and storage medium

Abstract

The application provides a wind farm reactive power coordination method and device and a storage medium, relates to the technical field of reactive power coordination, and specifically comprises the following steps: collecting environment parameters and power operation parameters of a wind farm, calculating a temperature and humidity adaptability index, a wind power benefit index, a load regulation index and a voltage stability index; combining the wind power benefit index, the load regulation index and the voltage stability index to generate an initial reactive power coordination index, correcting the initial reactive power coordination index by using the temperature and humidity adaptability index to obtain a comprehensive reactive power coordination index; collecting historical time series data of the wind farm, constructing a reactive power coordination model based on a deep learning algorithm, training the model by using the historical time series data, obtaining a trained reactive power coordination model, taking the comprehensive reactive power coordination index of the current wind farm as input, and adjusting reactive power according to a reactive power coordination strategy output by the model. The application improves the operation efficiency and stability of the wind farm in a dynamic power grid environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of reactive power coordination, in particular to a wind farm reactive power coordination method, device and storage medium. BACKGROUND

[0002] In the operation of wind farms, the reactive power coordination of the power system is a key factor to ensure the stability and safety of the power grid. Both the lack and excess of reactive power can lead to fluctuations in the voltage of the power grid, thereby affecting the normal operation of power equipment. Traditional reactive power regulation methods mostly use fixed algorithms and static models, relying on historical data for prediction and decision-making. This method often shows insufficient adaptability when facing rapidly changing weather conditions and power loads. For example, under extreme weather conditions, the rapid changes in environmental parameters such as wind speed, temperature and humidity can cause instability in power demand and generation capacity, and the traditional method cannot make timely adjustments, thereby causing safety hazards in the operation of the power grid. In addition, due to the lack of real-time monitoring and dynamic adjustment mechanisms, the traditional method can lead to overcompensation or insufficient compensation of reactive power, further exacerbating the burden on the power system.

[0003] The existing technology also has obvious deficiencies in the dynamic adjustment capability of the power system. Traditional reactive power coordination technology mainly relies on static models and empirical rules, lacking intelligent decision support based on real-time environmental data. This limits the response speed and flexibility of the power system when dealing with load fluctuations, equipment failures and other emergencies. At the same time, existing technology often fails to fully consider the mutual influence between different environmental factors, such as the combined effect of temperature and humidity on the power generation efficiency of wind farms, resulting in an inability to accurately assess the actual operating state of the wind farm. Therefore, a more intelligent and efficient solution is needed in the reactive power coordination of wind farms to improve the dynamic response capability and overall operating efficiency of the power system.

[0004] The above information disclosed in the background section is only intended to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] The purpose of the present application is to provide a wind farm reactive power coordination method, device and storage medium to solve the problems raised in the background technology.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The wind farm reactive power coordination method comprises the following specific steps:

[0008] Step 1: Collect the environmental parameters and power operation parameters of the wind farm in the current monitoring time period, the environmental parameters including wind speed, temperature and humidity, the power operation parameters including grid voltage, power generation power and load demand, first dimensionless processing is performed on the environmental parameters and power operation parameters, then the temperature and humidity are interactively processed to calculate the temperature and humidity adaptability index;

[0009] Step 2: Calculate the wind power benefit index according to the wind speed, power generation power and grid voltage, calculate the load adjustment index according to the load demand, power generation power and grid voltage, and calculate the voltage stability index according to the grid voltage and temperature;

[0010] Step 3: Combine the wind power benefit index, load adjustment index and voltage stability index to generate an initial reactive power coordination index, the weights of the wind power benefit index, load adjustment index and voltage stability index in the calculation of the initial reactive power coordination index are determined based on the analytic hierarchy process, and the initial reactive power coordination index is modified using the temperature and humidity adaptability index to obtain a modified comprehensive reactive power coordination index;

[0011] Step 4: Collect a plurality of sets of historical time series data of the wind farm in a plurality of historical monitoring time periods, including historical time series environmental parameters, historical time series power operation parameters and corresponding reactive power coordination strategies, construct a reactive power coordination model based on a deep learning algorithm, calculate the comprehensive reactive power coordination index of the wind farm in each historical time period as the input of the model, take the corresponding reactive power coordination strategy as the label, train the model to obtain the trained reactive power coordination model, take the comprehensive reactive power coordination index of the current wind farm as the input, adjust the reactive power according to the reactive power coordination strategy output by the model, and the reactive power coordination strategy is specifically the output change of the reactive power compensation device.

[0012] Further, the environmental parameters of the wind farm are collected, and the specific logic is that the wind speed, temperature and humidity of the wind farm are collected simultaneously, the collection frequency is once every minute, the average value of the meteorological information values in the monitoring time period is calculated, and the average value is taken as the meteorological information of the wind farm, the meteorological information including average wind speed, average temperature and average humidity; wherein the temperature refers to the atmospheric temperature at a height of 1.5-2 meters from the ground;

[0013] The average temperature and average humidity are interactively processed to calculate the temperature and humidity adaptability index, and the specific logic is as follows:

[0014] According to the following formula, the standard deviation of the temperature in the monitoring time period is calculated:

[0015] ;

[0016] In the formula, is the average temperature, is the standard deviation of the temperature, is the temperature collected at the i-th collection time in the monitoring time period, is the temperature collected at the i-th collection time in the monitoring time period, is the total number of collection times in the monitoring time period, is the index of the collection time in the monitoring time period;

[0017] The temperature fluctuation index is calculated according to the standard deviation of the temperature:

[0018] ;

[0019] In the formula, is the temperature fluctuation index, is the standard deviation of the temperature;

[0020] The standard deviation of the humidity in the monitoring time period is calculated according to the following formula:

[0021] ;

[0022] In the formula, is the average humidity, is the standard deviation of the humidity, is the humidity collected at the i-th collection time in the monitoring time period, is the total number of collection times in the monitoring time period, is the index of the collection time in the monitoring time period; The humidity fluctuation index is calculated according to the standard deviation of the humidity:

[0023]

[0024] ;

[0025] In the formula, is the humidity fluctuation index, is the standard deviation of the humidity;

[0026] The temperature-humidity adaptability index is calculated according to the temperature fluctuation index and the humidity fluctuation index:

[0027] ;

[0028] In the formula, is the temperature-humidity adaptability index, is the temperature fluctuation index, is the humidity fluctuation index.

[0029] Further, the wind power benefit index is calculated according to the average wind speed, the power generation and the grid voltage after dimensionless processing, and the formula is as follows:

[0030] ;

[0031] In the formula,​ is a wind power benefit index, is an average wind speed, is a wind speed reference threshold, is a grid voltage, is a voltage reference threshold, is a generated power, is a preset proportion coefficient, and ;

[0032] The load adjustment index is calculated according to the load demand, the generated power and the grid voltage after dimensionless processing, and the formula is as follows:

[0033] ;

[0034] In the formula, is a load adjustment index, is a load demand, is a generated power, is a grid voltage, represents the minimum voltage for safe operation of the grid, , is a preset proportion coefficient, , and satisfies ;

[0035] The voltage stability index is calculated according to the grid voltage and the average temperature after dimensionless processing, and the formula is as follows:

[0036] ;

[0037] In the formula, is a voltage stability index, is a standard voltage designed by the system, representing the normal working voltage of the grid, is a grid voltage, is a natural constant, is an average temperature, is a temperature reference threshold.

[0038] Further, the wind power benefit index, the load adjustment index and the voltage stability index are combined to generate an initial reactive power coordination index, and the weights of the wind power benefit index, the load adjustment index and the voltage stability index in the calculation of the initial reactive power coordination index are determined based on the analytic hierarchy process. The formula for generating the initial reactive power coordination index is as follows:

[0039] ;

[0040] In the formula, is an initial reactive power coordination index, is a wind power benefit index, is a load adjustment index, is a voltage stability index, , and are weights of the wind power benefit index, the load adjustment index and the voltage stability index respectively, which are determined according to the analytic hierarchy process.

[0041] Further, weights of the wind power benefit index, the load adjustment index and the voltage stability index in the initial reactive power coordination index calculation are determined according to the analytic hierarchy process, and the specific logic is as follows:

[0042] The three indexes of the wind power benefit index, the load adjustment index and the voltage stability index are marked, the values of relative importance between each other are determined through the nine-scale method, and a judgment matrix is constructed, wherein the wind power benefit index is marked as 1, the load adjustment index is marked as 2, and the voltage stability index is marked as 3, and the constructed judgment matrix is as follows:

[0043] ;

[0044] wherein, , all represent indexes of the indexes, and , , represents an importance degree of the index with the index of relative to the index with the index of , the importance degree adopts the 1-9 scale method, and the greater the value is, the greater the importance degree of the index with the index of relative to the index with the index of is, and , ;

[0045] Each element value in the judgment matrix is divided by the sum of the column to obtain a normalized judgment matrix after normalization, the mean value of each row element value in the normalized judgment matrix is calculated, the mean value of the first row element value is taken as the weight of the wind power benefit index, the mean value of the second row element value is taken as the weight of the load adjustment index, and the mean value of the third row element value is taken as the weight of the voltage stability index, the three weights are scaled in equal proportion with the constraint condition that the sum of the scaled values is equal to 1, and the scaled weights are taken as proportional coefficients of the corresponding indexes.

[0046] Further, the initial reactive power coordination index is modified using the temperature and humidity adaptability index to obtain a modified comprehensive reactive power coordination index.

[0047] ;

[0048] wherein, is the comprehensive reactive power coordination index, is an initial reactive power coordination index, is a temperature and humidity adaptability index, is a correction coefficient, used to adjust the degree of influence of the temperature and humidity adaptability index on the comprehensive reactive power coordination index.

[0049] Further, the specific logic for constructing the reactive power coordination model is as follows:

[0050] A plurality of sets of historical time series data of wind farms are collected and randomly divided into a training set and a test set. The historical time series data includes historical time series environmental parameters, historical time series power operation parameters, and corresponding reactive power coordination strategies. The corresponding comprehensive reactive power coordination index is calculated according to the historical time series data. A reactive power coordination model is constructed based on a deep learning algorithm. The comprehensive reactive power coordination index in the training set is used as input, and the corresponding reactive power coordination strategy is used as a label. The model is trained to obtain a trained reactive power coordination model. The comprehensive reactive power coordination index in the test set is substituted into the trained model to obtain a corresponding prediction result. The error between the prediction result and the actual value in the test set is calculated. It is determined whether the error meets a preset error threshold. If it meets, the trained model, i.e. the reactive power coordination model, is output. The comprehensive reactive power coordination index of the current wind farm is used as input, and the reactive power is adjusted according to the reactive power coordination strategy output by the model. If it does not meet, the training is continued.

[0051] The process of randomly dividing into a training set and a test set has the following specific logic: The historical time series data of the wind farm is randomly sorted. 80% of the sorted historical time series data is used as the training set, and the remaining 20% is used as the test set.

[0052] Further, the error is the mean absolute error, the root mean square error, and the determination coefficient of the prediction result and the actual value in the test set, which are as follows:

[0053] The mean absolute error is:

[0054]

[0055] The root mean square error is:

[0056]

[0057] The determination coefficient is:

[0058]

[0059] wherein, , and represent the actual value, the predicted value, and the average value of the reactive power coordination strategy corresponding to the data of the i-th test sample, respectively. ​​​​A group number of test samples in the test sample set.

[0060] The application further provides a wind farm reactive power coordination device.

[0061] The parameter acquisition module is configured to acquire environmental parameters and power operation parameters of the wind farm in a current monitoring time period, the environmental parameters including wind speed, temperature and humidity, and the power operation parameters including grid voltage, power generation power and load demand, and the environmental parameters and the power operation parameters are first subjected to dimensionless processing, and then the temperature and the humidity are subjected to interactive processing to calculate a temperature and humidity adaptability index.

[0062] The index calculation module is configured to calculate a wind power benefit index according to the wind speed, the power generation power and the grid voltage, calculate a load adjustment index according to the load demand, the power generation power and the grid voltage, and calculate a voltage stability index according to the grid voltage and the temperature.

[0063] The correction module is configured to combine the wind power benefit index, the load adjustment index and the voltage stability index to generate an initial reactive power coordination index, the weight of the wind power benefit index, the load adjustment index and the voltage stability index in the calculation of the initial reactive power coordination index being determined based on an analytic hierarchy process, and the initial reactive power coordination index being corrected using the temperature and humidity adaptability index to obtain a corrected comprehensive reactive power coordination index.

[0064] The model construction module is configured to acquire historical time series data of a plurality of groups of wind farms in a plurality of historical monitoring time periods, including historical time series environmental parameters, historical time series power operation parameters and corresponding reactive power coordination strategies, construct a reactive power coordination model based on a deep learning algorithm, calculate comprehensive reactive power coordination indexes of the wind farms in the historical time periods as inputs of the model, take the corresponding reactive power coordination strategies as labels, train the model to obtain a trained reactive power coordination model, take the comprehensive reactive power coordination index of the current wind farm as an input, adjust reactive power according to a reactive power coordination strategy output by the model, and the reactive power coordination strategy is specifically an output change of a reactive power compensation device.

[0065] The application further provides a computer readable storage medium having a computer program stored thereon, and the computer program is characterized in that the computer program can implement the wind farm reactive power coordination method when executed by a processor.

[0066] Compared with the prior art, the application has the following beneficial effects:

[0067] The present application can reflect the adaptability of the wind farm under different climate conditions in real time by non-dimensional processing of environmental parameters and calculation of temperature and humidity adaptability index. The introduction of this index enhances the sensitivity to environmental changes, making the adjustment of reactive power more accurate and timely. Secondly, the present application calculates the wind power benefit index, load adjustment index and voltage stability index, determines the weight by the analytic hierarchy process, and generates the initial reactive power coordination index. Not only the power generation capacity of the wind farm and the stability of the grid voltage are considered, but also the influence of load demand changes on reactive power is fully reflected, forming a comprehensive evaluation system. In addition, the reactive power coordination model constructed by using deep learning algorithm can effectively capture the complex relationship between reactive power coordination strategy and different operating states by training historical time series data, and realize accurate prediction of future reactive power demand. This adaptive adjustment mechanism greatly improves the operation efficiency and stability of the wind farm in the dynamic grid environment, ensuring the safe and reliable operation of the power system in the changing environment. BRIEF DESCRIPTION OF DRAWINGS

[0068] Figure 1 It is a whole method flowchart of the present application;

[0069] Figure 2 It is a parallel coordinate image of the load adjustment index, voltage stability index and initial reactive power coordination index;

[0070] Figure 3 It is a 3D bar image of the wind power benefit index and the initial reactive power coordination index;

[0071] Figure 4 It is a device module schematic diagram of the present application. DETAILED DESCRIPTION

[0072] In order to make the purpose, technical scheme and advantages of the present application clearer, the present application is further described in detail below combined with specific embodiments.

[0073] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0074] Example:

[0075] Please see Figures 1-3 The present invention provides a technical solution:

[0076] The reactive power coordination method for wind farms includes the following specific steps:

[0077] Step 1: Collect environmental parameters and power operation parameters of the wind farm during the current monitoring period. The environmental parameters include wind speed, temperature and humidity, and the power operation parameters include grid voltage, power generation and load demand. First, the environmental parameters and power operation parameters are dimensionless. Then, the temperature and humidity are interactively processed to calculate the temperature and humidity adaptability index.

[0078] In this embodiment, the environmental parameters of the wind farm are collected based on the following logic: wind speed, temperature and humidity are collected simultaneously at the wind farm at a frequency of once per minute. The average value of the meteorological information values ​​within the monitoring period is calculated and used as the meteorological information of the wind farm. The meteorological information includes average wind speed, average temperature and average humidity. Among them, temperature refers to the atmospheric temperature at a height of 1.5-2 meters above the ground.

[0079] The temperature and humidity adaptability index is calculated by interactively processing the average temperature and average humidity. The specific logic behind this calculation is as follows:

[0080] Calculate the standard deviation of temperature over the monitoring period using the following formula:

[0081] ;

[0082] In the formula, Average temperature The standard deviation of temperature, For the monitoring period, the first Temperatures collected at each sampling time the total number of collection time points in the monitoring time period, the index of the collection time point in the monitoring time period;

[0083] The temperature fluctuation index is calculated according to the standard deviation of temperature:

[0084] ;

[0085] wherein, is the temperature fluctuation index, used to represent the stability and fluctuation degree of temperature change, a high value indicates that the temperature change is small, and the temperature is relatively stable, is the standard deviation of temperature;

[0086] The formula establishes an inverse relationship between the temperature fluctuation index and the standard deviation . When decreases, the value increases, indicating that the temperature is more stable; the square term makes the change of the temperature fluctuation index more smooth. When the temperature fluctuation is small, the change of is not so sensitive; when the fluctuation increases, the will decrease rapidly. This smoothness helps in dynamic monitoring in practical applications, avoiding the misleading of sudden drastic changes in decision-making.

[0087] According to the following formula, the standard deviation of humidity in the monitoring time period is calculated:

[0088] ;

[0089] wherein, is the average humidity, is the standard deviation of humidity, is the humidity collected at the th collection time point in the monitoring time period, the total number of collection time points in the monitoring time period, the index of the collection time point in the monitoring time period;

[0090] The humidity fluctuation index is calculated according to the standard deviation of humidity:

[0091] ;

[0092] wherein, is the humidity fluctuation index, used to quantify the stability of humidity change, a high value indicates that the humidity change is small, and the humidity is relatively stable, is the standard deviation of humidity; ​​

[0093] This formula establishes an inverse relationship between the humidity fluctuation index and the standard deviation . When decreases, increases, indicating that the humidity is more stable; the square term makes the change of the humidity fluctuation index smoother. When the humidity fluctuation is small, the change of is not so sensitive; when the fluctuation increases, will decrease rapidly. This smoothness helps in dynamic monitoring in practical applications, avoiding the misleading of sudden drastic changes in decision-making.

[0094] According to the temperature fluctuation index and the humidity fluctuation index, the temperature-humidity adaptability index is calculated:

[0095] ;

[0096] where is the temperature-humidity adaptability index, which is used to comprehensively evaluate the adaptability of the system to temperature and humidity fluctuations, is the temperature fluctuation index, is the humidity fluctuation index. When increases, it indicates that the temperature change is relatively stable, and the adaptability of the system to temperature change improves, so correspondingly increases; similarly, when increases, increases accordingly; that is, there is a positive correlation between , and .

[0097] The use of the square term and can effectively emphasize the impact of large fluctuations on the adaptability index. Larger fluctuations will enhance their weight in the total in the form of squares, thus more significantly affecting the value of , and the square root processing makes the change of the adaptability index smoother; the formula, through the method of square sum, comprehensively considers the fluctuations of temperature and humidity, giving an overall adaptability evaluation. This method can effectively reflect the overall performance of the system when facing multiple environmental factors, rather than just the influence of a single factor.

[0098] The advantage of step 1 is that it systematically collects environmental parameters and power operation parameters of the wind farm and performs dimensionless processing to eliminate the influence of dimension on data analysis. This method specifically includes comprehensive monitoring of average wind speed, average temperature and average humidity, as well as real-time recording of grid voltage, power generation and load demand. This detailed parameter collection makes the data more accurate and reliable, providing a scientific basis for subsequent reactive power coordination adjustment. At the same time, the temperature and humidity adaptability index calculated after the interaction of average temperature and average humidity can comprehensively reflect the influence of environment on the operation of wind farm, enhancing the adaptability to climate change.

[0099] Compared with the prior art, step 1 has the advantages of higher real-time and accuracy. The traditional method often relies on static models and simplified parameter settings, which easily ignores the complex interaction of different environmental factors. By systematic collection and dynamic processing, this scheme can reflect the actual operation state and environmental adaptability of the wind farm in real time, thereby improving the response speed and accuracy of reactive power adjustment. This improvement not only optimizes the effect of reactive power coordination, but also lays a solid foundation for the implementation of the overall scheme, promoting the effective execution of subsequent steps and the accurate generation of comprehensive reactive power coordination index. Overall, the implementation of step 1 provides valuable data support for the subsequent training of deep learning model and the formulation of reactive power coordination strategy, significantly improving the operation efficiency of wind farm and the stability of power grid.

[0100] Step 2: Calculate the wind power benefit index according to the wind speed, power generation and grid voltage, calculate the load adjustment index according to the load demand, power generation and grid voltage, and calculate the voltage stability index according to the grid voltage and temperature;

[0101] In this embodiment, the wind power benefit index is calculated according to the dimensionless average wind speed, power generation and grid voltage, and the formula is as follows:

[0102] ;

[0103] In the formula, is the wind power benefit index, which quantifies the benefit and performance of the wind power system operation, and reflects the contribution of the current operating conditions to the wind power benefit. The higher the value, the better the current wind power system operating conditions; is the average wind speed, is the wind speed reference threshold, is the grid voltage, is the voltage reference threshold, is the power generation, is a preset proportion coefficient for adjusting the magnitude of the wind power benefit index, so that the numerical range of is suitable for actual analysis and evaluation, and , wherein the specific value is determined based on historical operation data and actual benefit level of the wind farm; wherein and P are the average values in the monitoring time period. The logarithmic function is used here to process the calculation of wind power benefit because the logarithmic function has a nonlinear characteristic, which makes the benefit grow at a gradually slowing speed with the increase of the input variable, which is more reasonable in practical application.

[0104] When increases, it means that the deviation of wind speed from the wind speed reference threshold increases, resulting in the decrease of overall power generation efficiency of the wind farm, decreases; when increases, it means that the deviation of grid voltage from the voltage reference threshold increases, in the case of excessively high or low voltage, the generator set cannot operate under the best conditions, thereby affecting the power generation power and the service life of the equipment, decreases; when increases, it means that the wind turbine is operating efficiently. Higher power output usually corresponds to better wind speed conditions and optimized operation of the system, which will help to improve the overall power generation efficiency, correspondingly increases; that is, it indicates that and are positively correlated, , and are negatively correlated.

[0105] The load adjustment index is calculated according to the dimensionless load demand, power generation power and grid voltage, and the formula is as follows:

[0106] ;

[0107] In the formula, is the load adjustment index, which is used to quantify the adjustment capacity of the power system under the conditions of load demand, power generation power and grid voltage. reflects the response ability of the power system to the load demand and the influence of the stability of the grid voltage on the operation of the system in a certain monitoring time period; is the load demand, is the power generation power, is the grid voltage, represents the minimum voltage for safe operation of the grid; wherein is the average value in the monitoring time period; , is a preset proportionality coefficient, , and satisfies ; this is because the load demand is the core factor of the operation of the power system, which directly affects the balance between supply and demand of the grid. If the load demand increases, but the power generation power cannot keep up, it may lead to instability of the grid, so the load demand is given a higher weight;

[0108] Partly reflects the matching of load demand and power generation, describes the size of the system load pressure, add a constant 1 in the denominator to avoid the denominator is zero resulting in meaningless formula, while small, reflect the higher load pressure; This part reflects the grid voltage deviation from the minimum safe operating voltage, describes the problem of grid operation stability; formula through the linear combination of two parts of the influencing factors, the load adjustment pressure and grid operation stability are considered. Load demand and power generation through the first item, grid voltage stability through the second item. This structure has strong physical intuition, and has certain calculation simplicity.

[0109] When increases, it means that more reactive power is needed to maintain the stability of the grid, so increases; power generation represent the actual power generation capacity of the wind farm, the higher the power generation, means to meet more load demand, so when increases, decreases; when the grid voltage is lower than the minimum voltage of safe operation, the system needs more reactive power to maintain stability, so when decreases, followed by increases; that is, and positive correlation, , and negative correlation.

[0110] According to the dimensionless treatment of the grid voltage and the average temperature, the voltage stability index is calculated, and the formula is as follows:

[0111] ;

[0112] In the formula, is the voltage stability index, which is used to quantify the stability level of the grid voltage under different operating conditions, reflects the voltage state of the field in real-time operation, and the indirect influence of environmental temperature fluctuation on voltage stability; is the standard voltage designed by the system, which represents the normal working voltage of the grid, is the grid voltage, is a natural constant, is the average temperature, is the temperature reference threshold.

[0113] When increases, indicating that the deviation between the actual grid voltage and the standard voltage increases, leading to a decrease in power quality and voltage instability, decreases, the voltage deviation is an important direct factor affecting the stability of the grid operation. Using the reciprocal form of the deviation can effectively amplify the impact of voltage deviation from the standard value, highlighting the negative impact on stability when the deviation is large; increases, indicating that the deviation between the temperature and the temperature reference threshold increases, which may lead to a decrease in the efficiency of the power system operation, an increase in the risk of equipment damage, and a decrease in power supply reliability, decreases. The influence of environmental temperature on the stability of the grid operation has a nonlinear characteristic: when the temperature deviation is small, the influence is weak, and when the deviation increases, the influence rapidly intensifies, and the use of an exponential function can reflect this nonlinear relationship. The product form of the formula reflects the comprehensive influence of grid voltage deviation and environmental temperature deviation on voltage stability. Voltage stability is directly affected by voltage deviation and indirectly affected by temperature deviation, and their effects are coupled rather than independent. Therefore, using the product form can more accurately describe the combined influence of the two on stability. That is, , and are negatively correlated.

[0114] The advantage of step 2 is to accurately calculate the wind power benefit index, load adjustment index, and voltage stability index, which can comprehensively reflect the operation efficiency, load adaptability, and grid stability of the wind farm. By combining the dimensionless environmental parameters and power operation parameters, this step can more accurately evaluate the performance of the wind farm under different loads and environmental conditions, providing a scientific basis for optimizing the adjustment strategy of reactive power.

[0115] Compared with the prior art, this method has the advantages of comprehensiveness and dynamics. Traditional adjustment methods often only focus on a single indicator, lacking comprehensive evaluation of multiple factors, resulting in one-sidedness of the adjustment strategy. Through the multi-dimensional analysis of step 2, the present invention can reflect the comprehensive state of the wind farm in a complex grid environment in real time, thereby effectively improving the accuracy and response speed of reactive power coordination. This comprehensive evaluation mechanism ensures that reactive power can be adjusted in time under conditions of load changes, climate fluctuations, etc., ensuring the safety and stability of the grid. In the overall scheme, the use of this step can promote the effective implementation of subsequent steps. The various indices generated provide a solid data basis for the subsequent initial reactive power coordination index calculation, making the final comprehensive reactive power coordination index more reliable, thereby optimizing the training effect of the reactive power coordination model and improving the overall operation efficiency of the wind farm and the stability of the grid.

[0116] Step 3: Combine the wind power benefit index, load regulation index and voltage stability index to generate an initial reactive power coordination index, and the weights of the wind power benefit index, load regulation index and voltage stability index in the calculation of the initial reactive power coordination index are determined based on the analytic hierarchy process, and the initial reactive power coordination index is modified using the temperature and humidity adaptability index to obtain a modified comprehensive reactive power coordination index;

[0117] In this embodiment, the wind power benefit index, load regulation index and voltage stability index are combined to generate an initial reactive power coordination index, and the weights of the wind power benefit index, load regulation index and voltage stability index in the calculation of the initial reactive power coordination index are determined based on the analytic hierarchy process, wherein the formula for generating the initial reactive power coordination index is:

[0118] ;

[0119] In the formula, is the initial reactive power coordination index, which is used to quantify the coordination state of reactive power in the power system, and the higher the value, the greater the required reactive power of the current wind farm; is the wind power benefit index, is the load regulation index, is the voltage stability index, , and are the weights of the wind power benefit index, load regulation index and voltage stability index respectively, which are determined according to the analytic hierarchy process. The non-linear characteristics of the exponential function can amplify small changes, which is suitable for the dynamic response of the power system, because small load regulation or voltage changes can have a significant impact on reactive power coordination.

[0120] A higher wind power benefit index means that the wind farm is efficient under current conditions and can effectively meet the load demand. In this case, the required reactive power output will be lower because the grid stability is better and the system can operate more efficiently. Therefore, the wind power benefit index is negatively correlated with the reactive power coordination index; when the load demand increases, the load regulation index usually increases, which means that more reactive power is needed to maintain the stability of the grid. At this time, the demand for reactive power coordination of the system also increases accordingly, so the load regulation index is positively correlated with the reactive power coordination index; if the voltage stability index is low, it means that the grid is more likely to be unstable under the current conditions, and more reactive power may be needed to improve the voltage level and ensure the safe operation of the system, so the voltage stability index is negatively correlated with the reactive power coordination index.

[0121] Table 1: Initial Reactive Power Coordination Index Statistics

[0122]

[0123] Based on the analysis of the 15 groups of data provided in Table 1, the following conclusions can be drawn:

[0124] Initial reactive power coordination index The trend of change clearly reflects the demand characteristics of wind farms for reactive power under different operating conditions. Analysis shows that the load regulation index is the most significant factor affecting value, and its increase directly leads to the rise of value, indicating that the demand for reactive power compensation is more urgent when the load demand increases or the grid voltage approaches the lower limit of safety. In contrast, the wind power benefit index and the voltage stability index show an inhibitory effect on value. When is high, it indicates that the wind farm is in a high-efficiency power generation state, and the system stability is good, thereby reducing the dependence on reactive power coordination; the increase of reflects that the grid voltage tends to be rated and the temperature influence is small, which also helps to reduce the value of Z, meaning that the system is more stable and the required reactive power support is correspondingly reduced. Through the weight coupling of the three, the dynamic balance relationship between power generation efficiency, load pressure and voltage stability is reflected.

[0125] In addition, the data further verifies that the index formula can effectively integrate multi-dimensional operating indicators to achieve quantitative evaluation of the reactive power coordination demand of wind farms. This model not only considers the independent influence of each factor, but also captures the interaction through a nonlinear function, providing reliable input features for subsequent deep learning-based reactive power coordination strategy generation, which helps to improve the adaptive adjustment capability of wind farms in complex operating environments and the stability of the power grid.

[0126] According to the analytic hierarchy process, the weights of wind power benefit index, load regulation index and voltage stability index in the calculation of initial reactive power coordination index are determined. The specific logic is as follows:

[0127] Mark the wind power benefit index, load regulation index and voltage stability index as three indicators, determine the relative importance values between each other through the nine-scale method, and construct the judgment matrix. Among them, the wind power benefit index is marked as 1, the load regulation index is marked as 2, and the voltage stability index is marked as 3. The constructed judgment matrix is:

[0128] ;

[0129] wherein , both represent the index of the index, and , , represent the index the index of the index of the greater the value, the greater the importance of the index of the index of the index of , ;

[0130] Divide each element value in the judgment matrix by the sum of its column to obtain a normalized judgment matrix, calculate the mean value of each row element value in the normalized judgment matrix, and take the mean value of the first row element value as the weight of the wind power benefit index, take the mean value of the second row element value as the weight of the load adjustment index, and take the mean value of the third row element value as the weight of the voltage stability index, and scale the three weights in equal proportion, and take the scaled weight as the proportional coefficient of the corresponding index.

[0131] The initial reactive power coordination index is modified using the temperature and humidity adaptability index to obtain a modified comprehensive reactive power coordination index.

[0132] ;

[0133] In the formula, is the comprehensive reactive power coordination index, which quantifies the coordination state of reactive power in the power system, and the greater the value, the greater the required reactive power under the current conditions; is the initial reactive power coordination index, is the temperature and humidity adaptability index, is a correction coefficient for adjusting the influence of the temperature and humidity adaptability index on the comprehensive reactive power coordination index, and . The formula combines the initial reactive power coordination index and the correction term in the form of a product, reflecting the weakening effect of environmental conditions on reactive power output.

[0134] When increases, it means that the adaptability of the environment to the power equipment increases, and the performance of the power equipment will be more stable, and the power system manager can consider running more loads under suitable environmental conditions, thereby reducing the demand for reactive power compensation equipment, so is negatively correlated with .

[0135] The advantage of step 3 is that by synthesizing the wind power benefit index, the load adjustment index and the voltage stability index, an initial reactive power coordination index is generated, and the weights of each index in the calculation are determined according to the analytic hierarchy process. This method not only improves the scientificity and rationality of the importance evaluation of each index in the reactive power coordination, but also makes the final comprehensive reactive power coordination index better reflect the actual operation condition of the wind farm by modifying the initial reactive power coordination index. This process ensures that the influence of each factor on the reactive power regulation is reasonably quantified, so as to achieve a more accurate regulation strategy.

[0136] Compared with the prior art, the advantage of this step is that the multi-dimensional comprehensive evaluation replaces the previous single or one-sided regulation strategy, avoiding the risk of unstable power grid or low operating efficiency due to neglecting a key factor. Traditional methods often fail to comprehensively consider the comprehensive performance of the wind farm under different operating states and environmental conditions. However, by introducing the analytic hierarchy process, this step quantifies the relative importance of different factors in the entire reactive power coordination process using a scientific methodology, greatly enhancing the adaptability and flexibility of reactive power coordination. In the overall scheme, this step can effectively provide a scientific and reliable basis for subsequent modification and optimization, making the output of the reactive power coordination model more targeted and effective. This systematic and comprehensive analysis method helps the wind farm to dynamically adjust the reactive power according to real-time environmental changes and power demand, thereby improving the overall operating efficiency of the wind farm and the stability of the power grid, and promoting better utilization of renewable energy.

[0137] Step 4: Collect several sets of historical time series data of wind farms in multiple historical monitoring time periods, including historical time series environmental parameters, historical time series power operation parameters and their corresponding reactive power coordination strategies. Based on a deep learning algorithm, a reactive power coordination model is constructed, the comprehensive reactive power coordination index of the wind farm in each historical time period is calculated as the input of the model, the corresponding reactive power coordination strategy is taken as the label, the model is trained, and the trained reactive power coordination model is obtained. The comprehensive reactive power coordination index of the current wind farm is taken as the input, and the reactive power is adjusted according to the output of the model. The reactive power coordination strategy is specifically the output change of the reactive power compensation device:

[0138] In this embodiment, the specific logic for constructing the reactive power coordination model is as follows:

[0139] The historical time series data of several groups of wind farms is collected and randomly divided into a training set and a test set. The historical time series data includes historical time series environmental parameters, historical time series power operation parameters, and corresponding reactive power coordination strategies. The historical monitoring time period with performance rising after adjustment of the reactive power coordination strategy is selected as the historical time series data for subsequent model training. The corresponding comprehensive reactive power coordination index is calculated according to the historical time series data. A reactive power coordination model is constructed based on a deep learning algorithm. The comprehensive reactive power coordination index in the training set is taken as the input, and the corresponding reactive power coordination strategy is taken as the label. The model is trained to obtain the trained reactive power coordination model. The comprehensive reactive power coordination index in the test set is substituted into the trained model to obtain the corresponding prediction result. The error between the prediction result and the actual value in the test set is calculated. It is judged whether the error meets the preset error threshold. If it meets, the trained model, i.e. the reactive power coordination model, is output. The comprehensive reactive power coordination index of the current wind farm is taken as the input. The reactive power is adjusted according to the reactive power coordination strategy output by the model. If it does not meet, the training is continued.

[0140] The process of randomly dividing into a training set and a test set has the following specific logic: The historical time series data of the wind farm is randomly sorted. 80% of the sorted historical time series data is taken as the training set, and the remaining 20% is taken as the test set.

[0141] The error is the mean absolute error, root mean square error and determination coefficient of the prediction result and the actual value in the test set, which is as follows:

[0142] The mean absolute error is:

[0143] ;

[0144] The root mean square error is:

[0145] ;

[0146] The determination coefficient is:

[0147] ;

[0148] Wherein, 、 and represent the actual value, the predicted value and the average value of the reactive power coordination strategy corresponding to the data of the first group of test samples. The number of groups of test samples in the test sample set is tested. The advantage of step 4 is that by collecting historical time series data and constructing a reactive power coordination model based on a deep learning algorithm, the comprehensive reactive power coordination index of the current wind farm is taken as input, thereby realizing accurate reactive power adjustment. This process not only utilizes the rich information in historical data to improve the accuracy of the model, but also realizes a data-driven method to dynamically optimize the reactive power coordination strategy, adapting to real-time changes in power demand and environmental conditions.

[0149] Compared with the prior art, the advantage of this step is that it fully utilizes the powerful capabilities of deep learning, which can automatically identify and learn complex nonlinear relationships, thereby providing higher prediction accuracy than traditional methods. Traditional methods often rely on experience or rules, making it difficult to fully adapt to complex and variable power systems. However, through repeated learning of historical data, this step makes the reactive power coordination strategy more flexible and intelligent, effectively improving the stability and reliability of the power grid. In the overall scheme, the use of this step can significantly enhance the regulation capability and response speed of the wind farm, making the adjustment of reactive power more timely and accurate. By implementing model-based dynamic decision-making, this scheme can effectively improve the overall benefits of wind farms in the power grid, promote the efficient use of renewable energy, and drive the development of green power. This process not only improves the economic benefits of wind farms, but also lays a solid foundation for the sustainable development of the power grid.

[0150] Please refer to Figure 4 , a wind farm reactive power coordination device, comprising:

[0151] A parameter acquisition module is configured to acquire environmental parameters and power operation parameters of the wind farm at a current monitoring time period, wherein the environmental parameters include wind speed, temperature, and humidity, and the power operation parameters include grid voltage, power generation power, and load demand. The environmental parameters and power operation parameters are first dimensionless processed, and then the temperature and humidity are interactively processed to calculate a temperature and humidity adaptability index.

[0152] An index calculation module is configured to calculate a wind power benefit index based on the wind speed, power generation power, and grid voltage, calculate a load adjustment index based on the load demand, power generation power, and grid voltage, and calculate a voltage stability index based on the grid voltage and temperature.

[0153] A correction module is configured to combine the wind power benefit index, load adjustment index, and voltage stability index to generate an initial reactive power coordination index, wherein the weights of the wind power benefit index, load adjustment index, and voltage stability index in the calculation of the initial reactive power coordination index are determined based on an analytic hierarchy process, and the initial reactive power coordination index is corrected using the temperature and humidity adaptability index to obtain a corrected comprehensive reactive power coordination index.

[0154] The model construction module is used for collecting historical time series data of a plurality of groups of wind power plants in a plurality of historical monitoring time periods, including historical time series environment parameters, historical time series power operation parameters and corresponding reactive power coordination strategies, constructing a reactive power coordination model based on a deep learning algorithm, calculating a comprehensive reactive power coordination index of the wind power plant in each historical time period as an input of the model, taking the corresponding reactive power coordination strategy as a label, training the model, obtaining a trained reactive power coordination model, taking the comprehensive reactive power coordination index of the current wind power plant as an input, and adjusting the reactive power according to the reactive power coordination strategy output by the model, and the reactive power coordination strategy is specifically an output change of a reactive power compensation device.

[0155] The application further comprises a computer readable storage medium, which stores a computer program, and the computer program can realize the wind power plant reactive power coordination method when executed by a processor.

[0156] The above formulas are all dimensionless values, and the formulas are obtained by collecting a large amount of data to simulate a formula of the nearest real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0157] The above embodiments can be realized by software, hardware, firmware or any combination thereof, in whole or in part. When realized by software, the above embodiments can be realized in the form of a computer program product in whole or in part. Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software methods depends on the specific application and design constraints of the technical solutions.

[0158] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, which can be located in one place or distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0159] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. A reactive power coordination method for wind farms, characterized in that, The specific steps include: Step 1: Collect environmental parameters and power operation parameters of the wind farm during the current monitoring period. The environmental parameters include wind speed, temperature and humidity, and the power operation parameters include grid voltage, power generation and load demand. First, the environmental parameters and power operation parameters are dimensionless. Then, the temperature and humidity are interactively processed to calculate the temperature and humidity adaptability index. Step 2: Calculate the wind power benefit index based on wind speed, power generation and grid voltage; calculate the load regulation index based on load demand, power generation and grid voltage; and calculate the voltage stability index based on grid voltage and temperature. Step 3: Combine the wind power benefit index, load regulation index and voltage stability index to generate the initial reactive power coordination index. The weights of the wind power benefit index, load regulation index and voltage stability index in the calculation of the initial reactive power coordination index are determined based on the analytic hierarchy process. The initial reactive power coordination index is then corrected using the temperature and humidity adaptability index to obtain the corrected comprehensive reactive power coordination index. Step 4: Collect several sets of historical time-series data of wind farms over multiple historical monitoring periods, including historical time-series environmental parameters, historical time-series power operation parameters, and their corresponding reactive power coordination strategies. Construct a reactive power coordination model based on a deep learning algorithm, calculate the comprehensive reactive power coordination index of the wind farm in each historical time period as the input of the model, use the corresponding reactive power coordination strategy as a label, train the model to obtain the trained reactive power coordination model, use the current comprehensive reactive power coordination index of the wind farm as input, and adjust the reactive power according to the reactive power coordination strategy output by the model. The reactive power coordination strategy is specifically the output change of the reactive power compensation equipment.

2. The wind farm reactive power coordination method according to claim 1, characterized in that: The specific logic for collecting environmental parameters of wind farms is as follows: wind speed, temperature, and humidity are collected simultaneously at the wind farm at a frequency of once per minute, and the monitoring period is 1 hour. The average value of meteorological information values ​​within the monitoring period is calculated, and this average value is used as the meteorological information of the wind farm. The meteorological information includes average wind speed, average temperature, and average humidity; among which, temperature refers to the atmospheric temperature at a height of 1.5-2 meters above the ground. The temperature and humidity adaptability index is calculated by interactively processing the average temperature and average humidity. The specific logic behind this is as follows: Calculate the standard deviation of temperature over the monitoring period using the following formula: ； In the formula, Average temperature, The standard deviation of temperature, For the monitoring period, the first Temperatures collected at each sampling time To monitor the total number of data collection points within the specified time period, This serves as an index for the collection times within the monitoring period; Calculate the temperature fluctuation index based on the standard deviation of temperature: ； In the formula, This is the temperature fluctuation index. The standard deviation of temperature; Calculate the standard deviation of humidity during the monitoring period using the following formula: ； In the formula, Average humidity, The standard deviation of humidity, For the monitoring period, the first Humidity collected at each sampling time To monitor the total number of data collection points within the specified time period, This serves as an index for the collection times within the monitoring period; Calculate the humidity fluctuation index based on the standard deviation of humidity: ； In the formula, The humidity fluctuation index. The standard deviation of humidity; Calculate the temperature and humidity adaptability index based on the temperature fluctuation index and humidity fluctuation index: ； in, The temperature and humidity adaptability index. This is the temperature fluctuation index. This is the humidity fluctuation index.

3. The reactive power coordination method for wind farms according to claim 2, characterized in that: The wind power benefit index is calculated based on the dimensionless average wind speed, power generation, and grid voltage, using the following formula: ； In the formula, For wind power benefit index, The average wind speed, This is the wind speed reference threshold. For grid voltage, This is the voltage reference threshold. For power generation capacity, The preset scaling factor, and ; The load regulation index is calculated based on the dimensionless load demand, power generation, and grid voltage, using the following formula: ； In the formula, The load adjustment index, To meet load demand, For power generation capacity, For grid voltage, Indicates the minimum voltage required for safe operation of the power grid. , This is a preset proportional coefficient. And satisfy ; The voltage stability index is calculated based on the dimensionless grid voltage and average temperature, using the following formula: ； In the formula, It is the voltage stability index. The standard voltage for system design represents the normal operating voltage of the power grid. This is the grid voltage. It is a natural constant. The average temperature. This is the temperature reference threshold.

4. The reactive power coordination method for wind farms according to claim 1, characterized in that: The initial reactive power coordination index is generated by combining the wind power efficiency index, load regulation index, and voltage stability index. The weights of these indices in the calculation of the initial reactive power coordination index are determined using the analytic hierarchy process (AHP). The formula used to generate the initial reactive power coordination index is as follows: ； In the formula, The initial reactive power coordination index, For wind power benefit index, The load adjustment index, It is the voltage stability index. , and The weights for the wind power benefit index, load regulation index, and voltage stability index are determined using the analytic hierarchy process (AHP).

5. The wind farm reactive power coordination method according to claim 4, characterized in that: The weights of the wind power efficiency index, load regulation index, and voltage stability index in the initial reactive power coordination index calculation are determined using the analytic hierarchy process (AHP). The specific logic behind this determination is as follows: The three indicators—wind power efficiency index, load regulation index, and voltage stability index—are labeled, and the relative importance between each pair is determined using the nine-scale method. A judgment matrix is ​​constructed, where the wind power efficiency index is labeled as 1, the load regulation index as 2, and the voltage stability index as 3. The constructed judgment matrix is ​​as follows: ； in, , Both represent the index of the index, and , , Indicates that the index is The exponent relative to the index is The importance index is used, with importance calculated using a 1-9 scale. The larger the value, the higher the index. The index is compared to the index. The greater the importance of the index, and , ; Divide each element value in the judgment matrix by the sum of its columns to obtain a normalized judgment matrix. Calculate the mean of the element values ​​in each row of the normalized judgment matrix. Use the mean of the first row as the weight of the wind power benefit index, the mean of the second row as the weight of the load regulation index, and the mean of the third row as the weight of the voltage stability index. With the constraint that the sum of the scaled values ​​equals 1, scale the three weights proportionally and use the scaled weights as the scaling coefficients of the corresponding indices.

6. The reactive power coordination method for wind farms according to claim 1, characterized in that: The initial reactive power coordination index is corrected using the temperature and humidity adaptability index to obtain the corrected comprehensive reactive power coordination index. ； In the formula, To achieve a comprehensive reactive power coordination index, The initial reactive power coordination index, The temperature and humidity adaptability index. This is a correction factor used to adjust the degree of influence of the temperature and humidity adaptability index on the comprehensive reactive power coordination index.

7. The reactive power coordination method for wind farms according to claim 1, characterized in that: The specific logic underlying the construction of the reactive power coordination model is as follows: Several sets of historical time-series data from wind farms are collected and randomly divided into training and testing sets. The historical time-series data includes historical environmental parameters, historical power operation parameters, and their corresponding reactive power coordination strategies. A comprehensive reactive power coordination index is calculated based on the historical time-series data. A reactive power coordination model is constructed based on a deep learning algorithm. The comprehensive reactive power coordination index from the training set is used as input, and the corresponding reactive power coordination strategy is used as a label to train the model, resulting in a trained reactive power coordination model. The comprehensive reactive power coordination index from the testing set is then substituted into the trained model to obtain the corresponding prediction results. The error between the prediction results and the actual values ​​in the testing set is calculated. It is determined whether the error meets a preset error threshold. If it does, the trained model, i.e., the reactive power coordination model, is output, and the reactive power is adjusted according to the reactive power coordination strategy output by the model, using the current comprehensive reactive power coordination index of the wind farm as input. If the threshold is not met, training continues. The process of randomly dividing the data into training and test sets is as follows: the historical time-series data of the wind farm is randomly sorted; 80% of the sorted historical time-series data is used as the training set, and the remaining 20% ​​is used as the test set.

8. The reactive power coordination method for wind farms according to claim 7, characterized in that: The errors are the mean absolute error, root mean square error, and coefficient of determination between the predicted results and the actual values ​​in the test set, as detailed below: The mean absolute error is: ； The root mean square error is: ； The coefficient of determination is: ； in, , and Representing the first The actual, predicted, and average values ​​of the reactive power coordination strategy corresponding to the data of the test sample group; This represents the number of test samples in the test sample set.

9. A reactive power coordination device for wind farms, characterized in that: The aforementioned wind farm reactive power coordination device is used to execute the wind farm reactive power coordination method according to any one of claims 1-8, comprising: The parameter acquisition module is used to collect environmental parameters and power operation parameters of the wind farm during the current monitoring period. The environmental parameters include wind speed, temperature and humidity, and the power operation parameters include grid voltage, power generation and load demand. First, the environmental parameters and power operation parameters are processed to be dimensionless, and then the temperature and humidity are processed interactively to calculate the temperature and humidity adaptability index. The index calculation module is used to calculate the wind power benefit index based on wind speed, power generation and grid voltage, the load regulation index based on load demand, power generation and grid voltage, and the voltage stability index based on grid voltage and temperature. The correction module is used to combine the wind power benefit index, load regulation index and voltage stability index to generate an initial reactive power coordination index. The weights of the wind power benefit index, load regulation index and voltage stability index in the calculation of the initial reactive power coordination index are determined based on the analytic hierarchy process. The temperature and humidity adaptability index is used to correct the initial reactive power coordination index to obtain the corrected comprehensive reactive power coordination index. The model building module is used to collect several sets of historical time-series data of wind farms in multiple historical monitoring periods, including historical time-series environmental parameters, historical time-series power operation parameters and their corresponding reactive power coordination strategies. Based on deep learning algorithms, a reactive power coordination model is built, and the comprehensive reactive power coordination index of the wind farm in each historical time period is calculated as the input of the model. The corresponding reactive power coordination strategy is used as a label to train the model, resulting in a trained reactive power coordination model. The current comprehensive reactive power coordination index of the wind farm is used as input, and the reactive power is adjusted according to the reactive power coordination strategy output by the model. The reactive power coordination strategy is specifically the output change of the reactive power compensation equipment.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it can implement the wind farm reactive power coordination method according to any one of claims 1-8.

Images