Grid connection stability regulation method for large-scale access of wind turbine cluster

Abstract

This invention relates to the field of wind power technology, specifically to a method for grid-connected stability control of large-scale wind turbine clusters. The method includes: acquiring real-time load data and wind power ratio data; obtaining a real-time load change significance coefficient and a wind power ratio change significance coefficient based on the real-time load data and the wind power ratio data, respectively; obtaining an inertia reduction coefficient based on the real-time load change significance coefficient and the wind power ratio change significance coefficient; obtaining the local area grid inertia change performance based on the inertia reduction coefficient and the real-time load data, thereby obtaining an inertia cumulative impact coefficient; adjusting the original weight matrix of an existing load prediction algorithm based on the inertia cumulative impact coefficient; and adjusting the load response capability based on the load prediction results of the load prediction algorithm. Using this invention, the operational stability of the power grid when wind power is connected can be significantly improved.

Description

Technical Field

[0001] This invention relates to the field of wind power technology, specifically to a method for grid connection stability control of large-scale wind turbine clusters. Background Technology

[0002] Wind power is characterized by volatility and uncertainty, which can have various adverse effects on the operation of wind turbines. Specifically, wind energy fluctuations directly lead to changes in the active power output of wind turbines, and may also cause reactive power fluctuations, which in turn cause changes in the turbine's terminal voltage. When wind turbines are connected to the grid, their reactive power demand may also negatively impact the stability of the voltage at the grid connection point.

[0003] Unlike traditional synchronous generators, wind turbines generally have weaker inertial response capabilities. With large-scale wind power integration into the grid, the overall inertia level of the system may decrease, posing a challenge to the dynamic stability of the power system. Existing control measures still have adverse effects on grid connection stability when addressing such issues, mainly due to real-time changes in both supply and demand, directly reflected in fluctuations in electrical parameters—particularly significant in local power grids. Wind power integration leads to a reduction in regional grid inertia, making operational fluctuations more likely. Therefore, it is necessary to conduct an in-depth analysis of the grid connection stability of wind turbine clusters under the background of grid inertia changes, and to establish an effective grid optimization scheduling and control mechanism accordingly. Summary of the Invention

[0004] To address the technical problem of poor grid stability caused by wind power grid connection, this invention aims to provide a grid connection stability control method for large-scale wind turbine cluster access. The specific technical solution adopted is as follows:

[0005] Obtain real-time load data and wind power ratio data;

[0006] Based on the real-time load data and the wind power ratio data, the real-time load change significance coefficient and the wind power ratio change significance coefficient are obtained respectively. The inertia reduction coefficient is obtained based on the real-time load change significance coefficient and the wind power ratio change significance coefficient.

[0007] Based on the inertia reduction coefficient and the real-time load data, the inertia variation performance of the local power grid is obtained, and then the inertia cumulative influence coefficient is obtained. The original weight matrix of the existing load prediction algorithm is adjusted according to the inertia cumulative influence coefficient, and the load response capability is adjusted according to the load prediction result of the load prediction algorithm.

[0008] Furthermore, the process of obtaining the significance coefficient of the change in the wind power ratio includes:

[0009] Calculate the wind power ratio derivative data, where the wind power ratio derivative data is the derivative of the wind power ratio data at each sampling time;

[0010] Obtain the wind power ratio derivative data at a preset sampling time and the wind power ratio derivative data at other sampling times within a preset time interval;

[0011] The difference between the wind power ratio derivative data at a preset sampling time and the wind power ratio derivative data at other sampling times within a preset time interval is determined as a first difference value, and the average value of the first difference value is determined as a first average value. Based on the first average value, the significance coefficient of the wind power ratio change corresponding to the preset sampling time within the preset time interval is obtained.

[0012] Furthermore, the process of obtaining the inertia reduction coefficient includes:

[0013] Within a preset range of variation in the time interval, the difference between the real-time load variation significance coefficient and the wind power ratio variation significance coefficient is obtained as a second difference value, and the average value of the second difference value is obtained as a second average value.

[0014] The inertia reduction coefficient is obtained by inversely limiting the second average value to a preset range.

[0015] Furthermore, the method also includes: when the real-time load change significance coefficient does not correspond to the wind power ratio change significance coefficient, using a spline interpolation algorithm to complete the real-time load change significance coefficient.

[0016] Furthermore, the process of obtaining the local area power grid inertia variation includes:

[0017] The variance of the inertia attenuation coefficient corresponding to the sampling time within the preset sampling period is calculated as the first variance, and the variance of the real-time load data is calculated as the second variance.

[0018] The local area power grid inertia variation is obtained based on the first variance and the second variance.

[0019] Furthermore, the process of obtaining the cumulative inertia influence coefficient includes:

[0020] The difference between the local area power grid inertia variation performance during the preset sampling period and the local area power grid inertia variation performance during the preset number of sampling periods before the preset sampling period is calculated as a third difference value, and the average value of the third difference value is calculated as a third average value.

[0021] The average value of the local area power grid inertia variation performance of the preset number of sampling periods before the preset sampling period is calculated as the fourth average value, and the difference between the local area power grid inertia variation performance of the preset sampling period and the fourth average value is calculated as the fourth difference value.

[0022] The inertia cumulative influence coefficient of the preset sampling period is obtained based on the third average value and the fourth difference.

[0023] Furthermore, adjusting the original weight matrix of the existing load prediction algorithm according to the inertia cumulative influence coefficient includes:

[0024] The cumulative inertia influence coefficient is used as the weight of the data sampled in the corresponding sampling period, and the influence weight matrix is ​​arranged.

[0025] The target weight matrix is ​​obtained by adjusting the original weight matrix of the existing load prediction algorithm based on the influence weight matrix.

[0026] The target weight matrix is ​​used as the final weight matrix of the existing load prediction algorithm.

[0027] Furthermore, the process of obtaining the target weight matrix includes:

[0028] The target weight matrix is ​​obtained by multiplying the influence weight matrix with the original weight matrix.

[0029] Furthermore, the real-time load data and the wind power ratio data are obtained through a SCADA system.

[0030] Furthermore, adjusting the load response capability based on the load prediction results of the load prediction algorithm includes:

[0031] Calculate the difference between the power value of the load prediction result and the power value at the current sampling time. If the difference is positive, instruct the energy storage system to discharge; if the difference is negative, instruct the energy storage system to charge.

[0032] The present invention has the following beneficial effects:

[0033] First, obtain real-time load data and wind power ratio data. This forms the data foundation for subsequent analysis.

[0034] Secondly, based on the real-time load data and the wind power ratio data, the significance coefficients of real-time load variation and wind power ratio variation are obtained respectively. Then, the inertia reduction coefficient is obtained based on these coefficients. The larger the value of the inertia reduction coefficient, the more significant the influence of the wind turbine cluster on the grid inertia.

[0035] Finally, based on the inertia reduction coefficient and the real-time load data, the inertia variation performance of the local power grid is obtained, and then the cumulative inertia impact coefficient is obtained. The original weight matrix of the existing load prediction algorithm is adjusted according to the cumulative inertia impact coefficient, and the load response capability is adjusted according to the load prediction results of the load prediction algorithm. This invention dynamically adjusts the weights of data in each period of the existing load prediction algorithm based on the inertia variation of the local power grid within the sampling period. This optimization can effectively reduce the control margin required by the system, thereby avoiding the adverse effects of weakened inertia response on the stability of wind power grid connection. Attached Figure Description

[0036] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 A flowchart of the grid connection stability control method for large-scale wind turbine cluster access provided in the first embodiment of the present invention;

[0038] Figure 2 A flowchart illustrating the process of obtaining the significance coefficient of wind power ratio variation provided in the second embodiment of the present invention;

[0039] Figure 3 A flowchart illustrating the process of obtaining the inertia reduction coefficient according to the third embodiment of the present invention;

[0040] Figure 4 A flowchart illustrating the process of obtaining the local area power grid inertia variation performance provided in the fourth embodiment of the present invention;

[0041] Figure 5 This is a flowchart illustrating the process of obtaining the cumulative inertia influence coefficient according to the fifth embodiment of the present invention. Detailed Implementation

[0042] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of the grid-connected stability control method for large-scale wind turbine cluster access proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0044] The specific scheme of the grid connection stability control method for large-scale wind turbine cluster access provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Please see Figure 1 The diagram illustrates a flowchart of a grid-connected stability control method for large-scale wind turbine cluster access provided in the first embodiment of the present invention, the method comprising:

[0046] S101. Obtain real-time load data and wind power ratio data.

[0047] The real-time load data and the wind power ratio data are acquired through a SCADA (Supervisory Control and Data Acquisition) system. The wind power ratio data refers to the proportion of wind power supplied by the grid in the actual power grid. Specifically, this ratio data refers to the power output ratio, that is, the percentage of wind power output in the total grid output power.

[0048] S102. Based on the real-time load data and the wind power ratio data, obtain the real-time load change significance coefficient and the wind power ratio change significance coefficient respectively, and obtain the inertia reduction coefficient according to the real-time load change significance coefficient and the wind power ratio change significance coefficient.

[0049] The process of obtaining the significance coefficient of the wind power ratio change will be described in detail in the second embodiment, and will not be repeated here.

[0050] The approach to obtaining the significance coefficient of the change in the proportion of wind power is as follows:

[0051] First, the analysis scope is defined, and the wind power ratio data is analyzed within a preset time interval. Second, the rate of change of the wind power ratio data within the preset time interval is analyzed, transforming the object of analysis from the wind power ratio data itself to its derivative, that is, from the data itself to the rate of change of the data. This is a prerequisite for assessing the significance of the "change". Then, the difference between the rate of change of the wind power ratio data at the preset sampling time and the rate of change of the wind power ratio data at other sampling times within the preset time interval is calculated, resulting in a set of difference values. Finally, these difference values ​​are averaged to form a single, comprehensive significance coefficient. This coefficient reflects the degree of deviation of the fluctuation rate at the preset sampling time from the average of other sampling times within the preset time interval. The larger the value, the more abnormal and significant the change at the preset sampling time.

[0052] The process for obtaining the real-time load variation significance coefficient is the same as the process for obtaining the wind power ratio variation significance coefficient, and will not be repeated here.

[0053] The approach to obtaining the significance coefficient of real-time load changes is as follows:

[0054] First, the analysis scope is defined, and the real-time load data is analyzed within a preset time interval. Second, the analysis focuses on the rate of change of the real-time load data within the preset time interval, transforming the object of analysis from the real-time load data itself to its derivative, i.e., changing the object of analysis from the data itself to the rate of change of the data. This is a prerequisite for assessing the significance of the "change". Then, by calculating the difference between the rate of change of the real-time load data at the preset sampling time and the rate of change of the real-time load data at other sampling times within the preset time interval, a set of difference values ​​is obtained. Finally, these difference values ​​are averaged to form a single, comprehensive significance coefficient. This coefficient reflects the degree of deviation of the fluctuation rate at the preset sampling time from the average of other sampling times within the preset time interval. The larger the value, the more abnormal and significant the change at the preset sampling time.

[0055] The process of obtaining the inertia reduction coefficient will be described in detail in the third embodiment, and will not be repeated here.

[0056] The method for obtaining the inertia reduction coefficient is as follows:

[0057] First, the difference between the real-time load variation significance coefficient and the wind power ratio variation significance coefficient is calculated to reflect the degree of system imbalance. Then, the difference is averaged over a given time interval to characterize the average imbalance level during that period. Finally, this average value is converted into an inertia reduction coefficient through inverse mapping—that is, the larger the difference, the smaller the coefficient, and the result is limited to a preset reasonable range.

[0058] S103. Based on the inertia reduction coefficient and the real-time load data, obtain the local area power grid inertia variation performance, and then obtain the inertia cumulative influence coefficient. Adjust the original weight matrix of the existing load prediction algorithm according to the inertia cumulative influence coefficient, and adjust the load response capability according to the load prediction result of the load prediction algorithm.

[0059] The process of obtaining the local area power grid inertia variation will be described in detail in the fourth embodiment, and will not be repeated here.

[0060] The approach to obtaining the local area power grid inertia variation is as follows:

[0061] First, the variances of the inertia attenuation coefficient and the real-time load data within the preset sampling period are calculated to quantify the system inertia state and the fluctuation intensity of the external load, respectively. Then, a comprehensive analysis is conducted based on these two variance indices to comprehensively evaluate the dynamic inertia response and stability performance of the local power grid under load disturbance.

[0062] The process of obtaining the cumulative inertia influence coefficient will be described in detail in the fifth embodiment, and will not be repeated here.

[0063] The method for obtaining the cumulative inertia influence coefficient is as follows:

[0064] First, the average difference between the local area's power grid inertia variation and recent historical data is calculated to measure its persistent trend. Second, the difference between the local area's power grid inertia variation and recent historical average is calculated to capture its instantaneous fluctuation amplitude. Finally, by combining these two indicators, which represent trend and volatility respectively, a final coefficient that can comprehensively reflect the cumulative effect of inertia influence is determined.

[0065] To improve the grid connection stability of local power grids, energy storage systems and grid-connected inverters are commonly used to enhance grid frequency response speed and shorten response time. To address the impact of intermittent fluctuations in wind turbine power generation, a tiered, multi-time-scale coordinated control strategy can be employed to effectively regulate grid connection stability.

[0066] Multi-timescale coordinated control can be divided into the following three levels:

[0067] Millisecond-level control (0–2 seconds): Millisecond-level response is achieved through rotor kinetic energy control, providing virtual inertia support and maintaining system frequency stability.

[0068] Minute-level control (2–30 seconds): Employing a converter coordination strategy, combined with virtual inertia and droop control, a single frequency adjustment is achieved to quickly smooth out power fluctuations.

[0069] Hourly control (over 30 minutes): Power prediction and scheduling decisions are made based on optimization algorithms using predictive models.

[0070] Analysis of the Inertia Influence in Coordinated Control:

[0071] In control at small timescales such as seconds and minutes, the impact of the inertia response of the local power grid is relatively limited; its impact is mainly reflected in hourly control processes. Therefore, based on the local power grid inertia variation data obtained during the sampling period, the weights of the corresponding sampled data can be dynamically adjusted in the prediction model, thereby reducing the control margin and effectively mitigating the adverse effects of weakened inertia response on the stability of wind power grid connection.

[0072] Adjusting the original weight matrix of the existing load prediction algorithm based on the inertia cumulative influence coefficient includes:

[0073] The cumulative inertia influence coefficient is used as the weight of the data sampled in the corresponding sampling period, and the influence weight matrix is ​​arranged.

[0074] The target weight matrix is ​​obtained by adjusting the original weight matrix of the existing load prediction algorithm based on the influence weight matrix.

[0075] The target weight matrix is ​​used as the final weight matrix of the existing load prediction algorithm.

[0076] Furthermore, the process of obtaining the target weight matrix includes:

[0077] The target weight matrix is ​​obtained by multiplying the influence weight matrix with the original weight matrix.

[0078] Furthermore, adjusting the load response capability based on the load prediction results of the load prediction algorithm includes:

[0079] Calculate the difference between the power value of the load prediction result and the power value at the current sampling time. If the difference is positive, instruct the energy storage system to discharge; if the difference is negative, instruct the energy storage system to charge.

[0080] The charging and discharging process of the energy storage system based on the aforementioned difference command is existing technology and will not be described in detail here.

[0081] With the large-scale grid connection of wind turbine clusters, the strong fluctuations in their power output will affect the grid operation characteristics of the connected area during the grid connection process. Although dynamic reactive power compensation and other methods are now widely used, the fundamental differences between wind turbines and traditional generators in terms of technical characteristics—primarily manifested in the weaker inertia response capability—mean that large-scale grid connection will still weaken the system inertia and cause time-series inertia fluctuations in the regional grid. Therefore, when assessing the actual stability of wind power grid connection, a specific distinction should be made based on the dynamic relationship of inertia on both the supply and demand sides, and effective grid optimization scheduling and control should be implemented accordingly.

[0082] The inertia response capability of wind turbines is far lower than that of traditional synchronous generators. Large-scale grid connection of wind turbines will lead to a decrease in the overall system inertia, thereby affecting the dynamic stability of the power grid. While existing control measures can alleviate grid stability issues to some extent, their effectiveness is essentially constrained by the dynamic changes on both the supply and demand sides, specifically manifested as real-time fluctuations in key electrical parameters. Especially in local power grids, the reduction in regional inertia caused by wind power integration is more significant and more likely to trigger grid operation fluctuations. Therefore, it is essential to conduct a refined analysis and response to the impact of inertia changes from the perspective of the interaction between supply and demand timelines.

[0083] Figure 2 The flowchart below shows the process of obtaining the significance coefficient of wind power ratio change provided in the second embodiment of the present invention. The process of obtaining the significance coefficient of wind power ratio change includes:

[0084] S201. Obtain the wind power ratio derivative data, wherein the wind power ratio derivative data is the derivative of the wind power ratio data at each sampling time.

[0085] S202. Obtain the wind power ratio derivative data at the preset sampling time and the wind power ratio derivative data at other sampling times within the preset time interval.

[0086] Using sampling time Several adjacent times constitute the sampling time time interval The time interval The value range is not fixed and can be set independently. The preferred number of sampling times in the time interval is 3 to 8.

[0087] When the time interval When the value is odd, it is based on the sampling time. The wind power proportional derivative data is centered here. For example, when the time interval... If the value is 5, then the wind power proportional derivative data within the time interval are respectively the sampling times. , , , as well as The wind power proportional derivative data, sampling time The wind power proportional derivative data is the first wind power proportional derivative data in the time interval, at the sampling time. The wind power ratio derivative data is the second wind power ratio derivative data in the time interval, and so on.

[0088] When the time interval When the value is even, the sampling time The number of data points on the left side of the wind power ratio derivative data can be one more than the number of data points on the right side, or vice versa. For example, when the time interval... If the value is 4, then the wind power proportional derivative data within the time interval are respectively the sampling times. , , as well as The wind power proportional derivative data or the wind power proportional derivative data within the time interval are respectively the sampling time. , , as well as The wind power proportional derivative data.

[0089] S203. Determine the difference between the wind power ratio derivative data at the preset sampling time and the wind power ratio derivative data at other sampling times within the preset time interval as the first difference value, and determine the average value of the first difference value as the first average value, and obtain the significance coefficient of the wind power ratio change corresponding to the preset sampling time within the preset time interval based on the first average value.

[0090] The significance coefficient of the change in the proportion of wind power can be expressed by the formula:

[0091] ;

[0092] Among them, the Indicates the sampling time The wind power proportional derivative data, the Indicates within the preset time interval Other sampling times The wind power proportional derivative data, subscript Used to distinguish between preset time intervals The wind power proportional derivative data at other sampling times are not included here. The wind power proportional derivative data, therefore the The value range is 1 to The Indicates within the preset time interval The preset sampling time mentioned above The corresponding significance coefficient of the change in the proportion of wind power, i.e. the first average value.

[0093] The The first difference is represented by the wind power ratio derivative data at a preset sampling time and the wind power ratio derivative data at any other sampling time within a preset time interval, wherein the first difference is a non-negative number.

[0094] Figure 3 The flowchart illustrates the process of obtaining the inertia reduction coefficient according to the third embodiment of the present invention. The process of obtaining the inertia reduction coefficient includes:

[0095] S301. Within a preset range of variation in the time interval, obtain the difference between the real-time load variation significance coefficient and the wind power ratio variation significance coefficient as a second difference value, and obtain the average value of the second difference value as a second average value.

[0096] The second average value can be expressed by the formula:

[0097] ;

[0098] Among them, the Indicates the time interval Below, sampling time The real-time load variation significance coefficient, the Indicates the time interval Below, sampling time The significance coefficient of the change in the proportion of wind power, time interval The preset variation range is from n to m, the This represents the absolute value, that is, the second difference. It is a non-negative value.

[0099] Furthermore, the method also includes: when the real-time load change significance coefficient does not correspond to the wind power ratio change significance coefficient, using a spline interpolation algorithm to complete the real-time load change significance coefficient.

[0100] When the time analysis range for real-time load data is set to be smaller than that for wind power ratio data, that is, when using the same statistical method (e.g., calculating the average, variance, etc.) for both real-time load data and wind power ratio data, but the time window lengths used for calculation are different, the mismatch in time scale (one data is dense, the other is sparse) will lead to deviations in calculation. By using interpolation techniques, sparse data can be "filled" into the time points of dense data, thereby eliminating the mismatch in time dimension and ensuring the accuracy of subsequent analysis.

[0101] S302. The second average value is reversed and restricted to a preset range to obtain the inertia reduction coefficient.

[0102] The process of obtaining the inertia reduction coefficient includes:

[0103] ;

[0104] Among them, the Indicates the sampling time The inertia attenuation coefficient, the This represents an exponential function with the natural constant e as its base.

[0105] Sampling time The larger the value of the inertia attenuation coefficient, the more significantly the electrical characteristics of the power grid at that sampling moment are affected by the wind turbine cluster.

[0106] The smaller the cumulative difference between the real-time load change significance coefficient and the wind power ratio change significance coefficient, the more similar the changes on both the supply and demand sides are, which means that the inertia of the local power grid is more affected by the access of wind turbine units.

[0107] After large-scale integration of wind power into local power grids, its inertial response will undergo a process from fluctuation to stability. The core of this impact lies in the temporal difference between the inertial responses on the supply and demand sides: when wind power accounts for a high proportion of local power generation and its own output fluctuates greatly, it will significantly weaken the inertial response of the regional power grid. Therefore, it is necessary to conduct an in-depth analysis of the inertial changes of the local power grid.

[0108] Figure 4 This is a flowchart illustrating the process of obtaining the local area power grid inertia variation performance according to the fourth embodiment of the present invention. The process of obtaining the local area power grid inertia variation performance includes:

[0109] S401. Calculate the variance of the inertia attenuation coefficient corresponding to the sampling time within the preset sampling period as the first variance and the variance of the real-time load data as the second variance.

[0110] The sampling period is in days.

[0111] S402. Obtain the local area power grid inertia variation performance based on the first variance and the second variance.

[0112] The variation in the inertia of the power grid in the local area can be expressed by the following formula:

[0113] ;

[0114] Among them, the Indicates the sampling period The first variance of the inertia attenuation coefficient corresponding to the sampling time, the Indicates the sampling period The second variance of the real-time load data corresponding to the sampling time within the range, the Indicates the sampling period The weighted measure of the first variance of the inertia attenuation coefficient at each sampling time shows a trend of changing from significant fluctuations to slow growth and then stabilizing as the sampling period number increases. This is the value obtained by calculating the logarithm to base 10 of the first variance. Indicates the sampling period The aforementioned local area power grid inertia variation behavior.

[0115] This describes the difference between the inertia attenuation that has occurred during the sampling period and the actual load fluctuation. In other words, it analyzes the current inertia performance of the local power grid based on the consistency of these two performances. A greater difference indicates that the local power grid is more significantly affected by energy storage control measures, meaning the inertia impact on the local power grid is greater. After weight adjustment (divided by) ), centralization (-1) and normalization ( After that, a standardized indicator is obtained. .

[0116] Therefore, the larger the value of the local power grid inertia variation, the more significant the impact of wind turbine cluster access. In this case, the stability control of the power grid becomes more difficult, especially under extreme weather conditions, where drastic changes in wind power output can easily lead to problems such as insufficient load supply.

[0117] Figure 5 The flowchart below shows the process for obtaining the cumulative inertia influence coefficient according to the fifth embodiment of the present invention. The process for obtaining the cumulative inertia influence coefficient includes:

[0118] S501. Calculate the difference between the local area power grid inertia variation performance in the preset sampling period and the local area power grid inertia variation performance in the preset number of sampling periods before the preset sampling period as the third difference value, and calculate the average value of the third difference value as the third average value.

[0119] The third difference quantity is a non-negative number, and the third difference quantity can be expressed as: The third average value can be expressed by the formula: , wherein Indicates the sampling period The aforementioned local area power grid inertia variation behavior, the Indicates the sampling period The aforementioned local area power grid inertia variation behavior, the The value range is 1 to The This indicates the preset number, such as setting it to 7.

[0120] S502. Calculate the average value of the local area power grid inertia variation performance of the preset number of sampling periods before the preset sampling period as the fourth average value, and calculate the difference between the local area power grid inertia variation performance of the preset sampling period and the fourth average value as the fourth difference quantity.

[0121] The fourth difference can be expressed by the formula: The Indicates the preset sampling period The fourth average value of the local area power grid inertia variation performance of the sampling period (including the preset sampling period) for a preset number of sampling periods (e.g., 5).

[0122] The This represents an absolute value function, meaning that the fourth difference quantity is non-negative.

[0123] S503. Obtain the inertia cumulative influence coefficient of the preset sampling period based on the third average value and the fourth difference.

[0124] The cumulative influence coefficient of inertia can be expressed by the formula:

[0125] ;

[0126] Among them, the Indicates the sampling period The inertia cumulative influence coefficient, the This represents a normalization function, such as the range normalization function.

[0127] The larger the value, the greater the cumulative effect coefficient of inertia. The greater the difference between the cumulative inertia impact coefficient and the commonly used one, the more likely it is to represent a new, non-random pattern of change, and therefore the more worthy of attention. The larger the value, the greater the cumulative effect coefficient of inertia. The value differs significantly from the average value of the cumulative inertia impact factor in the past, indicating that it is not stable or reliable enough. The larger the value, the greater the inertia influence of that sampling period, and the greater its role in subsequent inertia change prediction analysis. Therefore, the weight of this data should also be greater.

[0128] The present invention has the following beneficial effects:

[0129] First, obtain real-time load data and wind power ratio data. This forms the data foundation for subsequent analysis.

[0130] Secondly, based on the real-time load data and the wind power ratio data, the significance coefficients of real-time load variation and wind power ratio variation are obtained respectively. Then, the inertia reduction coefficient is obtained based on these coefficients. The larger the value of the inertia reduction coefficient, the more significant the influence of the wind turbine cluster on the grid inertia.

[0131] Finally, based on the inertia reduction coefficient and the real-time load data, the inertia variation performance of the local power grid is obtained, and then the cumulative inertia impact coefficient is obtained. The original weight matrix of the existing load forecasting algorithm is adjusted according to the cumulative inertia impact coefficient, and the load response capability is adjusted according to the load forecasting results of the load forecasting algorithm. Based on the inertia variation of the local power grid within the sampling period, the weights of the data in each period of the existing load forecasting algorithm can be dynamically adjusted. This optimization can effectively reduce the control margin required by the system, thereby avoiding the adverse effects of weakened inertia response on the stability of wind power grid connection.

[0132] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0133] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A method for grid connection stability control of large-scale wind turbine clusters, characterized in that, The method includes: Obtain real-time load data and wind power ratio data; Based on the real-time load data and the wind power ratio data, the real-time load change significance coefficient and the wind power ratio change significance coefficient are obtained respectively. The inertia reduction coefficient is obtained based on the real-time load change significance coefficient and the wind power ratio change significance coefficient. Based on the inertia reduction coefficient and the real-time load data, the inertia variation performance of the local power grid is obtained, and then the inertia cumulative influence coefficient is obtained. The original weight matrix of the existing load prediction algorithm is adjusted according to the inertia cumulative influence coefficient, and the load response capability is adjusted according to the load prediction result of the load prediction algorithm. The process of obtaining the local area power grid inertia variation performance includes: The variance of the inertia attenuation coefficient corresponding to the sampling time within the preset sampling period is calculated as the first variance, and the variance of the real-time load data is calculated as the second variance. The local area power grid inertia variation is obtained based on the first variance and the second variance. The process of obtaining the cumulative inertia influence coefficient includes: The difference between the local area power grid inertia variation performance during the preset sampling period and the local area power grid inertia variation performance during the preset number of sampling periods before the preset sampling period is calculated as a third difference value, and the average value of the third difference value is calculated as a third average value. The average value of the local area power grid inertia variation performance of the preset number of sampling periods before the preset sampling period is calculated as the fourth average value, and the difference between the local area power grid inertia variation performance of the preset sampling period and the fourth average value is calculated as the fourth difference value. The inertia cumulative influence coefficient of the preset sampling period is obtained based on the third average value and the fourth difference. Adjusting the original weight matrix of the existing load prediction algorithm based on the inertia cumulative influence coefficient includes: The cumulative inertia influence coefficient is used as the weight of the data sampled in the corresponding sampling period, and the influence weight matrix is ​​arranged. The target weight matrix is ​​obtained by adjusting the original weight matrix of the existing load prediction algorithm based on the influence weight matrix. The target weight matrix is ​​used as the final weight matrix of the existing load prediction algorithm.

2. The grid connection stability control method for large-scale wind turbine cluster access as described in claim 1, characterized in that, The process of obtaining the significance coefficient of the change in the proportion of wind power includes: Calculate the wind power ratio derivative data, where the wind power ratio derivative data is the derivative of the wind power ratio data at each sampling time; Obtain the wind power ratio derivative data at a preset sampling time and the wind power ratio derivative data at other sampling times within a preset time interval; The difference between the wind power ratio derivative data at a preset sampling time and the wind power ratio derivative data at other sampling times within a preset time interval is determined as a first difference value, and the average value of the first difference value is determined as a first average value. Based on the first average value, the significance coefficient of the wind power ratio change corresponding to the preset sampling time within the preset time interval is obtained.

3. The grid connection stability control method for large-scale wind turbine cluster access as described in claim 1, characterized in that, The process of obtaining the inertia reduction coefficient includes: Within a preset range of variation in the time interval, the difference between the real-time load variation significance coefficient and the wind power ratio variation significance coefficient is obtained as a second difference value, and the average value of the second difference value is obtained as a second average value. The inertia reduction coefficient is obtained by inversely limiting the second average value to a preset range.

4. The grid connection stability control method for large-scale wind turbine cluster access as described in claim 3, characterized in that, The method further includes: when the real-time load change significance coefficient does not correspond to the wind power ratio change significance coefficient, using a spline interpolation algorithm to complete the real-time load change significance coefficient.

5. The grid connection stability control method for large-scale wind turbine cluster access as described in claim 1, characterized in that, The process of obtaining the target weight matrix includes: The target weight matrix is ​​obtained by multiplying the influence weight matrix with the original weight matrix.

6. The grid connection stability control method for large-scale wind turbine cluster access as described in claim 1, characterized in that, The real-time load data and the wind power ratio data are obtained through the SCADA system.

7. The grid connection stability control method for large-scale wind turbine cluster access as described in claim 1, characterized in that, Adjusting load response capabilities based on the load prediction results of the load prediction algorithm includes: Calculate the difference between the power value of the load prediction result and the power value at the current sampling time. If the difference is positive, instruct the energy storage system to discharge; if the difference is negative, instruct the energy storage system to charge.

Images