New energy AGC credible frequency modulation power calculation method based on data driving

Abstract

The invention discloses a new energy AGC credible frequency modulation power calculation method based on data driving, and the method comprises the following steps: S1, obtaining a new energy field station frequency modulation standard reaching rate and a mapping relation between a reported frequency modulation capacity value and a fuzzy credible frequency modulation probability based on a historical operation data set, and building a frequency modulation capacity credibility function; s2, establishing an actual frequency modulation power optimization model of the new energy station based on a frequency modulation capacity credibility function, and constraining the actual frequency modulation power of each station within a feasible capability range; and S3, solving the actual frequency modulation power optimization model of the new energy station to obtain the actual frequency modulation power of each new energy electric field, and issuing the actual frequency modulation power to each new energy station. According to the new energy AGC frequency modulation power calculation method based on data driving provided by the invention, the problem of inaccurate frequency modulation capacity data caused by factors such as new energy output randomness and station capacity false report in the existing regulation and control process is avoided, and the operation safety of a power system is ensured.

Description

A Data-Driven Reliable Frequency Modulation Power Calculation Method for New Energy AGC Technical Field

[0001] This invention belongs to the field of new energy technology, specifically relating to a data-driven reliable frequency modulation power calculation method for new energy AGC. Background Technology

[0002] In new power systems, the large-scale grid connection of new energy sources such as wind and solar power poses a severe challenge to ensuring the stability and reliability of the power system. Automatic Generation Control (AGC) technology for new energy sources, which adjusts the output of new energy sources in real time to maintain the frequency stability of the power system, is a key technology for ensuring the stable operation of new power systems. However, due to the inherent randomness and volatility of new energy output, the frequency regulation capacity data reported by new energy power plants is often unreliable, making it difficult to accurately determine the control strategy for new energy AGC. Therefore, accurately determining the frequency regulation capacity of new energy sources is of great significance for improving the accuracy of new energy AGC control and ensuring the stable and reliable operation of new power systems.

[0003] Currently, traditional frequency regulation power calculation methods mainly allocate power proportionally based on the frequency regulation capacity values ​​reported by renewable energy power plants, and dynamically adjust them according to their actual response in the previous command cycle (Zhang Yanjun, Zhuang Hongshan, Li Yongguang, Yu Xiangtao, Ma Xiaolei, Wang Kaiwei. Design and application of multi-objective and multi-constraint automatic generation control system [J]. Electrical Engineering, 2019, (03): 12-16.). The specific adjustment mechanism is as follows: First, the frequency regulation response capability of the renewable energy power plant is determined by comparing the actual output change rate with a preset threshold. If the actual change rate reaches or exceeds the threshold, the power plant is considered to have the corresponding frequency regulation capability; otherwise, it is determined not to have it. For power plants determined not to have frequency regulation capability, their frequency regulation command values ​​will be redistributed to other power plants with frequency regulation capability. This method has a relatively one-sided evaluation process, based only on the response of a single historical cycle, and does not fully consider the random fluctuations and multi-cycle dynamic characteristics of renewable energy output. When the frequency regulation capacity reported by the power plant is inaccurate, the frequency security of the system will be difficult to guarantee. While traditional mathematical optimization methods have been applied to frequency regulation power allocation calculations, their modeling process often fails to adequately account for the stochastic fluctuations in renewable energy output (Zhang Lei, Ma Xiaowei, Wang Manliang, Chen Li, Gao Bingtuan. Distributed Cooperative Control Strategy for AGC Units in Interconnected Renewable Energy Power Systems [J]. China Electric Power, 2025, 58 (03): 8-19.). Similarly, when the frequency regulation capacity reported by renewable energy power plants is inaccurate, this limitation is mainly reflected in the boundary setting of constraints. The constraint boundaries established based on inaccurate frequency regulation capacity often deviate from the actual operating conditions, thus affecting the accuracy and practicality of the optimization results. Machine learning can train models based on a large amount of historical operating data, and mine the output characteristics and frequency regulation response patterns of renewable energy power plants through data-driven methods, thereby evaluating their frequency regulation capabilities. Its advantages lie in its ability to effectively handle complex nonlinear relationships and its strong adaptability, making it suitable for renewable energy power plants in different scenarios and of different types. However, its model interpretability is poor, and it has high requirements for professional technology and computing resources, making it difficult to implement in actual engineering. Against this backdrop, there is an urgent need to propose new methods for calculating the frequency regulation power of new energy sources in order to improve the accuracy of AGC control of new energy sources. Summary of the Invention

[0004] Based on the actual needs of new energy AGC regulation, this invention proposes a data-driven method for calculating the frequency regulation power of new energy AGC, which avoids the problem of inaccurate frequency regulation capacity data caused by factors such as the randomness of new energy output and false reporting of power station capacity in the existing regulation process, and ensures the safe operation of the power system.

[0005] The present invention is achieved by at least one of the following technical solutions.

[0006] A data-driven method for calculating the reliable frequency regulation power of new energy AGC includes the following steps: S1, obtaining the mapping relationship between the frequency regulation compliance rate of new energy power plants, the reported frequency regulation capacity value, and the fuzzy reliable frequency regulation probability based on historical operation datasets, and establishing a frequency regulation capacity reliability function; S2, establishing an optimization model for the actual frequency regulation power of new energy power plants based on the frequency regulation capacity reliability function, and constraining the actual frequency regulation power of each power plant to be within the feasible capability range; S3, solving the optimization model for the actual frequency regulation power of new energy power plants to obtain the actual frequency regulation power of each new energy power plant, and distributing it to each new energy power plant.

[0007] Furthermore, the frequency modulation capacity reliability function is:

[0008] In the formula, For the number of terms, As basis functions, The coefficients of the credibility function are... Here is the frequency modulation capacity reliability function. Let be the approximate reliability function of arbitrary polynomial chaotic expansion based on physical constraint embedding, c be the frequency modulation compliance rate, and r be the frequency modulation capacity reported by the station.

[0009] Furthermore, the basis functions are constructed based on the statistical moments of the data, and the sampled values ​​of the set frequency regulation compliance rate and the frequency regulation capacity values ​​reported by the new energy power stations are described by the tensor product of a set of univariate basis functions.

[0010] Furthermore, the coefficients of the confidence function are determined using the collocation method.

[0011] Furthermore, the optimization model for the actual frequency regulation power of new energy power plants takes minimizing the frequency regulation cost as its objective function.

[0012] Furthermore, the constraints of the actual frequency regulation power optimization model for new energy power plants include joint reliability constraints, system frequency regulation power balance constraints, system cross-sectional safety constraints, and unit ramping and slippage constraints. Among them, the joint reliability constraint requires that the product of the fuzzy credible frequency regulation probabilities of each new energy power plant is greater than or equal to the set total system confidence level; the system frequency regulation power balance constraint requires that the sum of the products of the frequency regulation command values ​​and the frequency regulation compliance rates of all new energy power plants must be equal to the total frequency regulation demand of the system; the system cross-sectional safety constraint requires that the actual capacity of the cross-section be less than the maximum allowable capacity value of the cross-section; and the unit ramping and slippage constraints require that the power change of each new energy power plant be less than the maximum ramping or slippage value.

[0013] Furthermore, the frequency regulation compliance rate is the ratio of the actual output value of the power station to the down-regulation command value within the corresponding control cycle at the end of the specified time.

[0014] Furthermore, the training data for the actual frequency regulation power optimization model of new energy power plants is a three-data pair consisting of the fuzzy reliable frequency regulation probability of historical operating data, the sampled value of the set frequency regulation compliance rate, and the frequency regulation capacity value reported by the new energy power plants.

[0015] A computer device includes a memory and a processor, the memory being electrically connected to the processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the method described above.

[0016] A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor implements the method.

[0017] Compared with existing technologies, the beneficial effects of this invention are as follows: (1) This invention proposes for the first time a data-driven frequency regulation capacity credibility function for new energy power stations. The credibility function is based on an arbitrary polynomial chaotic expansion method with physical constraint embedding. It uses historical data to establish a mapping relationship between the reported frequency regulation capacity value and the frequency regulation compliance rate of new energy power stations and the fuzzy credible frequency regulation probability, thereby realizing a quantitative assessment of the credibility of the frequency regulation capacity of new energy power stations and effectively quantifying the credibility of the frequency regulation capacity of new energy power stations.

[0018] (2) This invention proposes a method for calculating the reliable frequency regulation power of new energy AGC. Based on the reliability function established above, a new energy power plant frequency regulation power optimization model based on fuzzy reliable frequency regulation probability is proposed and solved. Unlike the traditional new energy AGC control model, this model transforms the upper and lower limit constraints of frequency regulation power into joint reliability constraints. When the frequency regulation capacity reported by the power plant is inaccurate, the frequency regulation power of each new energy power plant is calculated accurately and efficiently, and the frequency regulation power does not violate the actual upper and lower limit constraints.

[0019] (3) The present invention can accurately and efficiently calculate the frequency regulation power of new energy AGC, effectively reduce the frequency fluctuation risk caused by insufficient frequency regulation capability, improve the reliability and stability of the power grid, promote the consumption of renewable energy and the construction of new power systems, and bring good social and economic benefits. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as limiting the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 illustrates a data-driven reliable frequency modulation power calculation method for new energy AGC according to an embodiment of the present invention.

[0022] Figure 2 is a graph showing the total actual power versus the total commanded power curves at a certain cross section where the proposed method is applied in the embodiment.

[0023] Figure 3 is a frequency curve of a certain cross section in the embodiment where the proposed method is applied. Detailed Implementation

[0024] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0025] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, the technical solutions of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the specific embodiments described herein are only for explaining this application and are not intended to limit this application.

[0026] As shown in Figure 1, this embodiment of a data-driven reliable frequency regulation power calculation method for new energy AGC includes the following steps: S1, establishing an optimization model for the actual frequency regulation power of new energy power plants based on the frequency regulation capacity reliability function; S2, establishing constraints on the optimization model for the actual frequency regulation power of new energy power plants, constraining the actual frequency regulation power of each power plant to be within the feasible capability range; S3, solving the optimization model for the actual frequency regulation power of new energy power plants to obtain the actual frequency regulation power of each new energy power plant, and distributing it to each new energy power plant.

[0027] This invention proposes a credibility function to quantify the credibility of the frequency regulation capacity of renewable energy power plants. Its physical meaning is that, after obtaining the reported frequency regulation capacity of renewable energy power plants, it describes the probability of successful execution of the plant under different frequency regulation command compliance rates. This probability quantifies the credibility of the frequency regulation capacity of the renewable energy power plants. The essence of this method is not to predict renewable energy output, but to determine the credibility and degree of credibility of the reported frequency regulation capacity of renewable energy power plants based on historical data. The specific process is as follows: First, based on the historical operating data of renewable energy power plants, the reported frequency regulation capacity value, the sampled value of the frequency regulation compliance rate, and the fuzzy credible frequency regulation probability are calculated and statistically analyzed. Then, a credibility function based on arbitrary polynomial chaotic expansion with physical constraint embedding is proposed to determine the mapping relationship between the reported frequency regulation capacity value, the frequency regulation compliance rate, and the fuzzy credible frequency regulation probability. Based on the fuzzy credible frequency regulation probability index of each renewable energy power plant, the credibility of the frequency regulation capacity of the renewable energy power plant is quantified.

[0028] As one embodiment, the training data for the actual frequency regulation power optimization model of the new energy power station in this embodiment is a three-data pair consisting of the fuzzy reliable frequency regulation probability of historical operating data, the sampled value of the set frequency regulation compliance rate, and the frequency regulation capacity value reported by the new energy power station. Specifically, this includes: 1) Calculating discrete data pairs a) Frequency regulation compliance rate. The historical dataset of new energy power station a contains... Group operation data, including frequency regulation capacity values ​​reported by new energy power plants. Frequency modulation command value and the corresponding actual output value .

[0029] In the formula, These are the 1st, 2nd, ..., in the historical dataset. The frequency regulation capacity value reported by each new energy power station These are the 1st, 2nd, ..., in the historical dataset. Frequency regulation command value for new energy power stations These are the 1st, 2nd, ..., in the historical dataset. Actual output value of each new energy power station.

[0030] During the operation of renewable energy power plants, the plants must track the Automatic Generation Control (AGC) commands from the dispatch center in real time and achieve the required control accuracy within a specified response time to ensure the frequency stability of the power system. If the plant achieves the required control accuracy within the specified time, it is considered a qualified point; conversely, if the plant fails to achieve the required control accuracy within the specified time due to false reporting of frequency regulation capacity, changes in weather conditions, or other factors, it is considered a unqualified point. To quantify the frequency regulation performance of renewable energy power plants, this invention defines the frequency regulation compliance rate as the ratio of the plant's actual output value at the end of a specified time period to the value of the down-regulation command within that control cycle, i.e.: (1) In the formula, The frequency regulation compliance rate of the i-th historical data of the new energy power station. The actual output value of the i-th historical data of the new energy power station. The frequency modulation command value is the historical data of the i-th data point of the new energy power station. This represents the total historical data of new energy power stations.

[0031] b) Reliable Frequency Regulation Probability To quantitatively evaluate the performance of new energy power plants under different reported frequency regulation capacities and different frequency regulation compliance rates, this invention defines a reliable frequency regulation probability evaluation index. Specifically, the reliable frequency regulation probability refers to the ratio of the number of times the frequency regulation compliance rate is greater than or equal to a set compliance rate sampling value in the historical operating data corresponding to a given reported frequency regulation capacity value to the total number of all operating data corresponding to that reported frequency regulation capacity value. ,Right now: (2) Where the function Satisfies the following definition: (3) Where: The sampled value is the set frequency modulation compliance rate. This represents the total number of data points reported when the adjustment capacity is r.

[0032] The set frequency modulation compliance rate sampling value is derived from the frequency modulation compliance rate statistics in historical operating data. Specifically, firstly, a set of frequency modulation compliance rates is statistically analyzed from the historical operating data corresponding to the given reported frequency modulation capacity value. Then, the frequency modulation compliance rates are arranged in ascending order. Finally, from... Initially, the gradient values ​​are accumulated sequentially. When the accumulated gradient sum G reaches or exceeds the threshold T, the current data point is selected. As a sample value of the FM compliance rate, and resetting the cumulative sum, until all data points have been traversed, the resulting set of FM compliance rate sample values ​​is as follows: (4) The formula for calculating the gradient cumulative sum G is as follows: (5) In the formula, For the k-th frequency modulation compliance rate sample value, It is a data value in the historical data that is greater than the i-th frequency modulation compliance rate.

[0033] c) Fuzzy Confidential Frequency Modulation Probability: As an example, the above-mentioned frequency modulation task compliance determination is based on whether the actual frequency modulation compliance rate reaches or exceeds the set sampling value. It is a binary determination mechanism, i.e., compliance is marked as 1, and non-compliance is marked as 0. However, this determination method has significant limitations in practical applications, mainly in that it lacks a progressive assessment of the frequency modulation task completion status, making it difficult to capture subtle differences when the frequency modulation compliance rate is close to but has not fully reached the preset value, thus failing to comprehensively reflect the actual degree of completion of the frequency modulation task.

[0034] Therefore, to more accurately assess the performance of new energy power plants under different frequency regulation compliance conditions, this invention proposes using fuzzy membership functions to describe the degree of frequency regulation task completion. Specifically, the input to the fuzzy membership function is the difference between the actual frequency regulation compliance rate and the sampled value of the frequency regulation compliance rate, and the output is the membership degree of task completion (the degree of task completion). By establishing fuzzy membership functions, continuous quantitative evaluation of the frequency regulation task completion can be achieved, thereby providing a more accurate quantitative basis for the reliability assessment of the frequency regulation capacity of new energy power plants.

[0035] This invention selects a Z-shaped membership function based on expert experience to describe the degree of frequency regulation task completion. First, for a given reported frequency regulation capacity value, the data set consisting of the difference between the actual frequency regulation compliance rate and the sampled frequency regulation compliance rate is presented to an expert review panel composed of grid dispatchers, new energy power plant operation and maintenance engineers, and power system analysis experts. The expert review panel selects a membership degree for each data point. Then, the Z-shaped function structure in equation (6) is selected as the structure of the membership function. Finally, the specific membership function can be determined by least squares.

[0036] (6) In the formula, Membership function; and The parameters are the Z-shaped function parameters, where .

[0037] Furthermore, this invention defines an evaluation index for fuzzy reliable frequency modulation probability, which is the ratio of the sum of the membership degrees of each completed frequency modulation task in the historical operational data corresponding to a given reported frequency modulation capacity to the total number of all operational data corresponding to that reported frequency modulation capacity value. ,Right now: In equation (7), For a given reported frequency modulation capacity, Let be the frequency regulation compliance rate in the j-th group of historical operating data corresponding to the given reported frequency regulation capacity. To report the frequency modulation capacity as given The corresponding amount of historical running data.

[0038] Fuzzy reliable frequency regulation probability can accurately reflect the degree to which current renewable energy power plants complete frequency regulation tasks. If the fuzzy reliable frequency regulation probability of each frequency regulation task can be predicted, the true frequency regulation capacity of renewable energy power plants can be accurately obtained, thereby significantly improving the accuracy and security of power grid frequency control.

[0039] In summary, the fuzzy reliable frequency regulation probability of each sample operation data is calculated using equation (7), and combined with the sampled value of the set frequency regulation compliance rate and the frequency regulation capacity value reported by the new energy power station to form a three-data pair. There are A pairs of sample data as the original training data for the mapping.

[0040] 2) Credibility function of arbitrary polynomial chaotic expansion based on physical constraint embedding.

[0041] This invention proposes a credibility function based on arbitrary polynomial chaotic expansion with embedded physical constraints to obtain an explicit mapping function between a set frequency regulation compliance rate, the frequency regulation capacity value reported by new energy power plants, and the fuzzy credible frequency regulation probability. First, the functional form based on arbitrary polynomial chaotic expansion is given; then, the basis functions of the function are constructed from the statistical moments of the data; finally, the coefficients in the credibility function are determined using the collocation method incorporating physical constraints.

[0042] a) Mapping Function Based on Arbitrary Polynomial Chaotic Expansion: The credibility function of this invention is essentially a mapping model. Its inputs are the frequency regulation compliance rate and the frequency regulation capacity value reported by new energy power plants, and its output is the fuzzy credibility frequency regulation probability. Arbitrary polynomial chaotic expansion is based on a data-driven approach, utilizing the moment information (such as mean, variance, and higher-order moments) of the input variables themselves to construct orthogonal polynomial basis functions. These orthogonal polynomial basis functions establish a mapping model between the input and output data. Since the basis functions are constructed based on the moment information of the input variables themselves, the arbitrary polynomial chaotic expansion polynomial can capture the main influence of the input on the output with the fewest terms, thus achieving high accuracy and convergence speed.

[0043] In this invention, for any given frequency regulation compliance rate sample value and the frequency regulation capacity value reported by the renewable energy power station, the corresponding fuzzy reliable frequency regulation probability can be obtained through a p-order mapping function based on an arbitrary polynomial chaotic expansion. To approximate, that is: (8) In the formula, As basis functions, The coefficients of the credibility function are... Here is the frequency modulation capacity reliability function. Let c be the approximate reliability function of an arbitrary polynomial chaotic expansion based on physical constraint embedding, where c is the frequency modulation compliance rate, r is the frequency modulation capacity reported by the power station, and M is the number of terms, the value of which can be determined by the dimension n and the highest order p. The specific calculation formula is as follows: (9) The mapping model is constructed based on sample data. Next, this invention will explain how to determine the basis functions. and the coefficients of the polynomial .

[0044] b) Constructing basis functions from statistical moments of the data .

[0045] Since the frequency regulation compliance rate sample value set in the input data is independent of the frequency regulation capacity value reported by the new energy power stations, its basis functions can be described by the tensor product of a set of univariate bases: (10) Where: Let f(x) represent a set of basis functions, where the order of the basis functions is denoted by f(x). From 0 to Values, For variables The order basis functions, Let represent another set of basis functions, where the order of the basis functions is denoted by . From 0 to Values, For the first variable r order basis functions, This indicates the calculation of the tensor product.

[0046] For the sake of brevity, and Through respectively and The formula for its calculation is as follows: (11) (12) Where: For variables The q-th order basis function corresponds to the coefficients with a power of 0. For variables The q-th basis function corresponds to the coefficient with a power of 0.

[0047] in addition, and The polynomial orthogonality condition should be satisfied, that is: (13) (14) Due to and Since the polynomial orthogonality conditions of equations (11) and (12) are satisfied, two systems of equations can be derived: (15) (16) Where: For variables The coefficients of power functions with a power of 0 in the 0th order basis functions. For variables The coefficients of power functions with a power of 0 in the first-order basis functions. For variables The coefficients of power functions with a power of 1 in the first-order basis functions. For variables The coefficients of power functions with a power of 0 in the 0th order basis functions. For variables The coefficients of power functions with a power of 0 in the first-order basis functions. For variables The coefficients of power functions with a power of 1 in the first-order basis functions. For variables The distribution space, For variables The distribution space.

[0048] In the first equation of the system of equations (15) Substituting the first equation into the second equation yields a new equation, which is then substituted into the third equation, and so on. Thus, the system of equations (13) can be simplified to: (17) Where: For the sample data of( The first moment can be approximated by a sample of size A in the sample data: (18) Equation (17) can be written in matrix form as follows: (19) Where: For the sample data of Step moment, For variables of The corresponding powers in the basis functions are The coefficients of the power function can be determined by solving equation (19). That will confirm .

[0049] Similarly, in the first equation of the system of equations (16) Substituting the first equation into the second equation, we obtain a new equation. Then, we substitute this new equation into the third equation, and so on. Thus, the system of equations (14) can be simplified to: (20) Where: For the sample data of( The first moment can be approximated by a sample of size A in the sample data: (21) Equation (20) can be written in matrix form as follows: (22) By solving equation (22), the coefficients can be determined. That will confirm .

[0050] Therefore, in determining and Then, basis functions It can be obtained through equation (8).

[0051] c) Determining coefficients using the collocation method incorporating physical constraints .

[0052] Based on the physical laws governing the output of new energy power plants, the fuzzy reliable frequency modulation probability decreases as the sampled value of the frequency modulation compliance rate increases. Therefore, the first derivative of the function should be non-positive, and the fuzzy reliable frequency modulation probability value should belong to the range [0,1]. Thus, the mapping function needs to satisfy the following physical constraint: (twenty three)

[0053] (24) Therefore, in determining the polynomial coefficients When doing so, physical constraints need to be considered. Calculation coefficients. It can be expressed as a least-squares regression problem with a finite number of interpolation points and physical constraints.

[0054] First, construct the inequality constraint matrix equation. The derivative constraint can be expressed as follows: (25) Where: is the derivative matrix, and its specific form is shown in equation (26); The coefficient matrix is ​​a polynomial, as shown in equation (25).

[0055] (26) (27) of which is the polynomial coefficient in formula (23).

[0056] The upper and lower limit constraints can be expressed as follows: (28) Where: The basis function matrix has the following specific form: (29) Then, considering constraints (25) and (28), construct a quadratic programming problem: (30) Where: Let be the value matrix of the basis functions at the sample points, and its specific form is shown in equation (31); The response matrix is ​​shown in Equation (30).

[0057] (31) (32) Where: basis functions At point The function value, Let A be the fuzzy reliable frequency modulation probability of the Ath sample data.

[0058] Finally, the quadratic programming problem (30) can be solved using a commercial solver, thus obtaining the polynomial coefficients. .

[0059] In summary, the basis functions have been determined. and polynomial coefficients Subsequently, the credibility function of the arbitrary polynomial chaotic expansion based on physical constraint embedding can be determined. By inputting the frequency regulation compliance rate and the frequency regulation capacity value reported by the renewable energy power plants, the corresponding fuzzy credibility frequency regulation probability can be mapped.

[0060] 3) Calculation of the actual frequency regulation power of renewable energy power plants: The power system dispatching department calculates the frequency regulation power command based on the frequency regulation capacity reported by the renewable energy power plants and adjusts their frequency regulation output in real time to ensure system frequency security. However, due to factors such as the possibility of overstated capacity reporting by renewable energy power plants and uncertainties in meteorological conditions, the reported frequency regulation capacity often has deviations. This can lead to dispatching commands issued by the dispatching department violating the upper and lower limits of the actual frequency regulation power, thus compromising system frequency security.

[0061] To overcome the aforementioned shortcomings, this invention proposes a frequency regulation power optimization model for new energy power plants based on fuzzy reliable frequency regulation probability. Unlike traditional new energy automatic generation control and regulation models that directly embed the upper and lower limits of frequency regulation power as hard inequality constraints, this invention cleverly transforms the upper and lower limit constraints of frequency regulation power into a joint reliability constraint based on the fuzzy reliable frequency regulation probability index. Specifically, this model indirectly and effectively constrains the actual frequency regulation power of each power plant within its feasible capacity range by constraining the lower limit of the probability that the entire system can successfully complete the frequency regulation task. This modeling method cleverly avoids situations where the calculated frequency regulation power violates the actual upper and lower limit constraints when the frequency regulation capacity reported by the new energy power plant is inaccurate.

[0062] Establish an optimization model for the actual frequency regulation power of renewable energy power plants based on fuzzy reliable frequency regulation probability. The model takes minimizing frequency regulation cost as its objective function. (33) Among them, This represents the total number of new energy power plants. Indicates the first The unit frequency regulation capacity cost of a new energy power plant For the first Frequency regulation compliance rate of new energy power plants For the first Frequency regulation command value for each new energy electric field.

[0063] The constraints include joint reliability constraints, system frequency regulation power balance constraints, system cross-sectional safety constraints, and unit ramp-up and ramp-down constraints. Among them, the joint reliability constraint requires that the product of the fuzzy credible frequency regulation probabilities of each new energy power plant be greater than or equal to the set total system confidence level, as shown in equation (34). Its physical meaning is that the probability that all new energy power plants participating in frequency regulation in the entire power system can complete the frequency regulation task cannot be lower than the set value.

[0064] (34) Among them, This represents the overall confidence level of the system.

[0065] The system frequency regulation power balance constraint requires that the sum of the products of the frequency regulation command values ​​and the frequency regulation compliance rates of all new energy power plants must equal the total frequency regulation demand of the system, as shown in equation (35). Its physical meaning is that the system's frequency regulation power and frequency regulation demand must be balanced.

[0066] (35) Among them, This is for the overall frequency regulation requirements of the system.

[0067] The system cross-sectional safety constraint needs to ensure that the actual capacity of the cross-section is less than the maximum allowable capacity value of the cross-section, as shown in equation (36). Its physical meaning is to ensure that the capacity of the cross-section will not exceed the limit during system operation.

[0068] (36) Among them, and These represent the upper and lower limits of power flow transmission at section s, respectively. This represents the power transfer distribution factor of the new energy electric field at section s; and These represent the forward and reverse tidal current relaxation variables of section s, respectively.

[0069] The ramp-up and landslide constraints of the generating units need to ensure that the power change of each new energy power station is less than the maximum ramp-up or landslide value, as shown in equation (37). Its physical meaning is to ensure that the power change of the new energy power station meets the actual physical constraints.

[0070] (37) Among them, Indicates the station in the previous control period The actual contribution of the individual; and They represent the stations The upper limit of the slope and the upper limit of the landslide.

[0071] The established optimization model for the actual frequency regulation power of new energy power plants is a nonlinear optimization problem. By calling commercial solvers such as CPLEX or Gurobi, the frequency regulation power values ​​of each new energy power plant can be accurately solved.

[0072] The method proposed in this invention calculates the frequency regulation power of new energy AGC based on a data-driven credibility function, which can avoid the problem of inaccurate frequency regulation capacity data caused by factors such as the randomness of new energy output and false reporting of station capacity in the existing frequency regulation process, thereby ensuring the security of system frequency.

[0073] The above embodiments of the present invention are merely examples for clearly illustrating the present invention and are not intended to limit the implementation of the present invention. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively describe all possible implementations here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A data-driven reliable frequency modulation power calculation method for new energy AGC, characterized in that, Includes the following steps: S1. Based on historical operation datasets, obtain the mapping relationship between the frequency regulation compliance rate of new energy power plants, the reported frequency regulation capacity value, and the fuzzy reliable frequency regulation probability, and establish a frequency regulation capacity reliability function; S2. Based on the frequency regulation capacity reliability function, establish an optimization model for the actual frequency regulation power of new energy power plants, and constrain the actual frequency regulation power of each power plant to be within the feasible capability range; S3. Solve the optimization model for the actual frequency regulation power of new energy power plants to obtain the actual frequency regulation power of each new energy power plant, and distribute it to each new energy power plant.

2. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 1, characterized in that, The frequency modulation capacity reliability function is: In the formula, For the number of terms, As basis functions, The coefficients of the credibility function are... Here is the frequency modulation capacity reliability function. Let be the approximate reliability function of arbitrary polynomial chaotic expansion based on physical constraint embedding, c be the frequency modulation compliance rate, and r be the frequency modulation capacity reported by the station.

3. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 2, characterized in that, The basis functions are constructed based on the statistical moments of the data, and the tensor product of a set of univariate bases is used to describe the sampled value of the set frequency regulation compliance rate and the frequency regulation capacity value reported by the new energy power station.

4. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 2, characterized in that, The coefficients of the confidence function are determined by the collocation method.

5. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 1, characterized in that, The optimization model for actual frequency regulation power of new energy power plants takes minimizing frequency regulation cost as its objective function.

6. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 1, characterized in that, The constraints of the actual frequency regulation power optimization model for new energy power plants include joint reliability constraints, system frequency regulation power balance constraints, system cross-sectional safety constraints, and unit ramping and slope constraints. Among them, the joint reliability constraint requires that the product of the fuzzy credible frequency regulation probabilities of each new energy power plant is greater than or equal to the set total system confidence level; the system frequency regulation power balance constraint requires that the sum of the products of the frequency regulation command values ​​and the frequency regulation compliance rates of all new energy power plants must be equal to the total frequency regulation demand of the system; the system cross-sectional safety constraint requires that the actual capacity of the cross-section is less than the maximum allowable capacity value of the cross-section; and the unit ramping and slope constraints require that the power change of each new energy power plant is less than the maximum ramping or slope value.

7. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 6, characterized in that, The frequency regulation compliance rate is the ratio of the actual output value of the power station to the down-regulation command value within the corresponding control cycle at the end of the specified time.

8. The data-driven reliable frequency modulation power calculation method for new energy AGC according to claim 1, characterized in that, The training data for the optimization model of actual frequency regulation power of new energy power plants is a three-data pair consisting of the fuzzy reliable frequency regulation probability of historical operating data, the sampled value of the set frequency regulation compliance rate, and the frequency regulation capacity value reported by the new energy power plants.

9. A computer device comprising a memory and a processor, the memory being electrically connected to the processor, the memory storing a computer program, characterized in that: When the computer program is executed by the processor, it causes the processor to implement the method as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor implements the method as described in any one of claims 1 to 8.

Images