A method and system for predicting energy storage frequency modulation commands based on the radius method

Abstract

The present invention discloses a method and system for predicting energy storage frequency modulation commands based on the radius method, belonging to the technical field of power system dispatching and optimization, including: evaluating the prediction difficulty value of the frequency modulation command sequence through the radius method; improving the ANFIS prediction network structure according to the prediction difficulty value; predicting the value of the frequency modulation command through the improved ANFIS network and outputting the result. The present invention solves the problems existing in the existing frequency modulation command prediction methods, such as the lack of evaluation of the prediction difficulty of the nonlinearity of the original data, the inability to perform personalized model adjustment for different sequence complexities, and the delay in response time, and improves the prediction accuracy.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching and optimization, and particularly to a method and system for predicting energy storage frequency modulation commands based on the radius method. Background Art

[0002] With the rapid development of renewable energy and energy storage technologies, frequency modulation command prediction technology has gradually become an important part of smart grid management. The traditional method of using hybrid energy storage (ultracapacitor + lithium battery) to assist thermal power units in frequency modulation mainly relies on the instantaneous regulation ability of thermal power units. However, with the increasing proportion of new energy, the frequency modulation demand has become more complex. To address this challenge, researchers have proposed various auxiliary energy storage methods, including hybrid energy storage systems (such as the combination of ultracapacitors and lithium batteries) to respond more efficiently to grid frequency fluctuations. This technology can not only provide rapid energy response but also maximize the economic benefits of power plants. In recent years, data-driven prediction models, such as artificial neural networks (ANN), adaptive neuro-fuzzy inference systems (ANFIS), etc., have been increasingly widely used in frequency modulation prediction, laying a foundation for integrating various energy forms in the future.

[0003] Although the existing technologies have made certain progress in frequency modulation command prediction, there are still multiple deficiencies. First, traditional prediction methods often lack an assessment of the prediction difficulty in the original data, resulting in poor response caused by time delay or signal transmission delay that is not easily grasped. Second, the original sequence of frequency modulation commands usually has highly nonlinear and difficult-to-regulate characteristics, which makes it impossible for some prediction models with unified parameter tuning to obtain ideal prediction results. More importantly, traditional methods fail to achieve personalized customization for the complexity of specific frequency modulation sequences, resulting in different prediction effects and being unable to adapt to frequency modulation demands of different complexities. These problems not only affect the real-time response ability of energy storage devices but also further weaken the economic benefits of power plants. Therefore, there is an urgent need for a method to evaluate the prediction difficulty and optimize the prediction model structure accordingly to improve the prediction accuracy and response speed of frequency modulation commands, thereby ensuring the stability and economy of the power grid. Summary of the Invention

[0004] To solve the above technical problems, a method for predicting energy storage frequency modulation commands based on the radius method is proposed, including evaluating the prediction difficulty value of the frequency modulation command sequence by the radius method; improving the ANFIS prediction network structure according to the prediction difficulty value; predicting the value of the frequency modulation command through the improved ANFIS network and outputting the result; the prediction difficulty value includes a prediction difficulty evaluation, expressed as:

[0005] ,

[0006] Among them, D represents the prediction difficulty value, K represents the number of centers of two similar circles that are satisfied simultaneously, and N represents the number of times of collecting frequency modulation commands.

[0007] As a preferred embodiment of the method for predicting energy storage frequency modulation commands based on the radius method according to the present invention, wherein: the frequency modulation command sequence includes, assuming the frequency modulation command is , expressed as:

[0008] ,

[0009] Among them, represents the value of the frequency modulation command, , respectively represent different collection times; the sampling interval of the frequency modulation command is 1 second, and the coordinate value of in the rectangular coordinate system is .

[0010] As a preferred embodiment of the method for predicting energy storage frequency modulation commands based on the radius method according to the present invention, wherein: evaluating the prediction difficulty value of the frequency modulation command sequence by the radius method includes determining the center position and radius length of two adjacent points in the frequency modulation command sequence in the rectangular coordinate system; assuming the two adjacent points are and respectively, and the coordinate values of the two points in the rectangular coordinate system are and respectively. Connect A and B, and make the perpendicular bisector of line segment AB. Then the center of the two adjacent points is on the line , and the line is expressed as:

[0011] ,

[0012] Assume the center of the two adjacent points is , is always moving downward on the center line , and a circle is formed with the distance from to as the radius. By dynamically adjusting the center position, line segment AB divides the circle into two parts S1 and S2. During the process of forming the circle, when S1 / S2 < 0.03, the radius length is , and the center is .

[0013] As a preferred embodiment of the method for predicting energy storage frequency modulation commands based on the radius method according to the present invention, wherein: evaluating the prediction difficulty value of the frequency modulation command sequence by the radius method further includes obtaining the radius sequence of pairwise adjacent points and the center sequence in the frequency modulation command sequence by the radius method ;

[0014] Perform similarity analysis on the radius and center of all adjacent points in the frequency modulation instruction sequence to find the points with similar center coordinates and similar radius; the points with similar center coordinates include calculating the distance between two points in the center sequence, and setting the farthest distance to , set a center The distance from the center of other circles is less than / 2, then As the center of the circle, / Draw a circle with the radius / The center of the circle with the same radius is defined as the center of the circle with similar coordinates. Let the number be Radius similarity includes setting the radius sequence The radius of the circle with the largest radius is , the minimum radius is , the average value is , if the radius If the similarity interval is satisfied, then the radius in the similarity interval satisfies the radius similarity. Let the number be Similar intervals are expressed as:

[0015] ,

[0016] If a circle center satisfies both center similarity and radius similarity, then the number of similarities that are satisfied at the same time is recorded as .

[0017] As a preferred solution of the method for predicting energy storage frequency modulation instructions based on the radius method described in the present invention, the prediction difficulty value also includes that the larger the value of D, the greater the difficulty of the prediction.

[0018] As a preferred solution of the method for predicting energy storage frequency modulation instructions based on the radius method described in the present invention, wherein: the improving the ANFIS prediction network structure according to the prediction difficulty value includes adjusting the membership function of the ANFIS network based on the prediction difficulty value D and ; According to the difficulty of prediction, the membership function Improve and adjust the membership function , expressed as:

[0019] ,

[0020] in, 、 represent the width and center of the Gaussian function, respectively. Indicates the membership function based on the prediction difficulty The adjusted and improved value is expressed as:

[0021] ,

[0022] The membership function is improved and adjusted according to the prediction difficulty After the improvement and adjustment, the membership function is expressed as:

[0023] ,

[0024] Wherein, represents the value of the membership function adjusted and improved according to the prediction difficulty is expressed as:

[0025] ,

[0026] and The maximum value is 1 and the minimum value is 0.

[0027] As a preferred scheme of a method for predicting energy storage frequency modulation commands based on the radius method according to the present invention, wherein: predicting the value of the frequency modulation command and outputting the result includes inputting the frequency modulation command sequence into the improved ANFIS network for prediction. The improved ANFIS network dynamically adjusts the network structure according to the prediction difficulty value D and outputs the prediction result; using the prediction result as the final value for the energy storage system to act in advance.

[0028] Another object of the present invention is to provide a system for predicting energy storage frequency modulation commands based on the radius method. The present invention solves the problems existing in the existing frequency modulation command prediction methods, such as the lack of evaluation of the prediction difficulty of the nonlinearity of the original data, the inability to perform personalized model adjustment for different sequence complexities, and the delay in the response time, and improves the prediction accuracy. By carefully analyzing the original frequency modulation sequence, evaluating its prediction difficulty, and dynamically adjusting the structure of the ANFIS network according to the evaluation result, the efficient prediction of the frequency modulation command and the dynamic adjustment of the prediction network structure are realized. While reducing the response time of the prediction network and improving the prediction accuracy, it also improves the frequency modulation response speed and economic benefits of the power plant.

[0029] As a preferred scheme of a system for predicting energy storage frequency modulation commands based on the radius method according to the present invention, it is characterized in that it includes a frequency modulation evaluation module, a difficulty network module, and a prediction output module; the frequency modulation evaluation module is used to evaluate the prediction difficulty value of the frequency modulation command sequence by the radius method; the difficulty network module is used to improve the ANFIS prediction network structure according to the prediction difficulty value; the prediction output module is used to predict the value of the frequency modulation command by the improved ANFIS network and output the result.

[0030] A computer device includes a memory and a processor. The memory stores a computer program. It is characterized in that when the processor executes the computer program, the steps of the method for predicting energy storage frequency modulation instructions based on the radius method are implemented.

[0031] A computer-readable storage medium stores a computer program thereon. It is characterized in that when the computer program is executed by a processor, the steps of the method for predicting energy storage frequency modulation instructions based on the radius method are implemented.

[0032] Advantages of the present invention: The method for predicting energy storage frequency modulation instructions based on the radius method provided by the present invention quantifies and analyzes the difficulty of frequency modulation instructions through the prediction difficulty evaluation based on the radius method, can effectively identify the complexity of the frequency modulation instruction sequence, reduce the prediction error caused by data nonlinearity and complexity, and thus improve the accuracy of the prediction result; adjusts the network structure of ANFIS through the prediction difficulty value, enables the prediction model to be flexibly and personalized customized according to frequency modulation instructions of different complexities, ensures that the model can adapt to the changing frequency modulation instructions, and thus further improves the prediction effect; predicts the value of the frequency modulation instruction in advance through the ANFIS prediction network, enables the energy storage system to act in advance, reduces the time delay caused by signal transmission and equipment response, and the introduction of advance can improve the response speed of the energy storage device, optimize the frequency modulation performance of the power plant, and thus improve the economic benefit; secondly, due to the improvement of the prediction accuracy and response speed, the dependence and cost on standby power generation equipment can be reduced, the frequency modulation effect can be increased, and the frequency modulation cost can be reduced. Description of the Drawings

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for description in the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can also obtain other accompanying drawings without creative efforts based on these drawings.

[0034] Figure 1 It is the overall flowchart of the method for predicting energy storage frequency modulation instructions based on the radius method provided by an embodiment of the present invention.

[0035] Figure 2 It is a schematic diagram of the method for determining the radius length of the method for predicting energy storage frequency modulation instructions based on the radius method provided by an embodiment of the present invention.

[0036] Figure 3 It is the system scheme module diagram of the system for predicting energy storage frequency modulation instructions based on the radius method provided by an embodiment of the present invention. Detailed Embodiments

[0037] To make the above objects, features, and advantages of the present invention more apparent and understandable, the following provides a detailed description of the specific embodiments of the present invention in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0038] Example 1. Refer to Figure 1 - Figure 2 , which is the first embodiment of the present invention. This embodiment provides a method for predicting energy storage frequency modulation commands based on the radius method, including:

[0039] S1: Evaluate the prediction difficulty value of the frequency modulation command sequence through the radius method.

[0040] Furthermore, the frequency modulation command sequence includes. Let the frequency modulation command be , which is expressed as:

[0041] ,

[0042] where represents the value of the frequency modulation command, , respectively represent different acquisition times; the sampling interval of the frequency modulation command is 1 second, and the coordinate value of the value in the rectangular coordinate system is .

[0043] It should be noted that evaluating the prediction difficulty value of the frequency modulation command sequence through the radius method includes determining the center position and radius length of two adjacent points in the frequency modulation command sequence in the rectangular coordinate system; let the two adjacent points be and , and the coordinate values of the two points in the rectangular coordinate system are and respectively. Connect A and B, and make the perpendicular bisector of line segment AB. Then the center of the two adjacent points is on the line , and the line is expressed as:

[0044] ,

[0045] As Figure 2 shown, let the center of the two adjacent points be , has been moving downward on the center line , and with to A circle is formed with a radius equal to the distance. By dynamically adjusting the position of the center of the circle, line segment AB divides the circle into two parts, S1 and S2. During the process of forming the circle, for the areas of S1 and S2, when S1 / S2 < 0.03, the radius length is , and the center of the circle is .

[0046] It should also be noted that evaluating the prediction difficulty value of the frequency modulation instruction sequence by the radius method also includes obtaining the radius sequence of adjacent points in pairs in the frequency modulation instruction sequence and the center sequence ; performing similarity analysis on the radii and centers of all adjacent points in the frequency modulation instruction sequence to find points with similar center coordinates and similar radii; for points with similar center coordinates, it includes calculating the distance between two points in the center sequence. Let the maximum distance be , and assume a center . If the distance from this center to other centers is less than / 2, then a circle is drawn with as the center and as the radius. The centers that fall within the circle with a radius of / are defined as centers with similar coordinates. Let the number be / ; for similar radii, it includes assuming that the radius of the circle with the largest radius in the radius sequence is , the minimum radius is , the average value is , . If the radius satisfies the similarity interval, then the radii within the similarity interval satisfy radius similarity. Let the number be ; the similarity interval is expressed as:

[0047] ,

[0048] If a center satisfies both center similarity and radius similarity, then the number that satisfies both similarities is denoted as .

[0049] It should also be noted that the prediction difficulty value also includes prediction difficulty evaluation, which is expressed as:

[0050] ,

[0051] where D represents the prediction difficulty value; the larger the value of D, the greater the prediction difficulty.

[0052] It should also be noted that the center of the circle moves downward along the perpendicular bisector, and a circle is drawn with the distance from the center of the circle to the point as the radius until the area S1 / S2 of the two parts S1 and S2 divided by the line segment AB in the process of forming the circle is less than 0.03, and the radius length is determined. This process ensures that the shape of the circle matches the distribution characteristics of the data points by dynamically adjusting the radius, thereby reflecting the local fluctuation of the frequency modulation instruction sequence; by determining the radius length and the position of the center of the circle, a quantitative standard is provided for the subsequent evaluation of the prediction difficulty of the frequency modulation instruction sequence. The smaller the radius length, the more intense the fluctuation between the two adjacent points and the greater the prediction difficulty; conversely, the larger the radius length, the smoother the change of the data points and the lower the prediction difficulty, which achieves the accurate characterization of the local characteristics of the frequency modulation instruction sequence and lays the foundation for the quantitative evaluation of the prediction difficulty.

[0053] S2: Improve the ANFIS prediction network structure according to the prediction difficulty value.

[0054] Furthermore, improving the ANFIS prediction network structure according to the prediction difficulty value includes adjusting the membership function of the ANFIS network based on the prediction difficulty value D. and ; According to the difficulty of prediction, the membership function Improve and adjust the membership function , expressed as:

[0055] ,

[0056] in, 、 represent the width and center of the Gaussian function respectively, Indicates the membership function based on the prediction difficulty Adjust the improved value, expressed as:

[0057] ,

[0058] According to the difficulty of prediction, the membership function Improve and adjust the membership function , expressed as:

[0059] ,

[0060] in, Indicates the membership function based on the prediction difficulty Adjust the improved value, expressed as:

[0061] ,

[0062] and The maximum value is 1 and the minimum value is 0.

[0063] It should be noted that the main fuzzy inference system FIS consists of five functional modules, including a rule base, a database, a decision-making unit, a fuzzy interface, and a defuzzification result. The initial membership functions and rules are defined according to the actual situation. An adaptive network is a multi-layer feedforward network composed of square nodes and circular nodes. The nodes are connected together through directed links. The directional nodes have parameters, while the circular nodes do not. The adaptive neuro-fuzzy inference system ANFIS is composed of the combination of FIS and an adaptive network. By adjusting the fuzzy inference rules and membership functions, the performance of FIS can be improved. Moreover, ANFIS combines the interpretability characteristics of FIS and the learning ability of the adaptive network in terms of function. ANFIS consists of two parts. The first part is the antecedent, and the second part is the consequent, which are interconnected in the form of a network through rules.

[0064] Taking three inputs and nine rules as an example for illustration, the ANFIS model structure is divided into five layers, and each layer contains several nodes. The first layer is the fuzzification layer, which performs a fuzzification process representing the membership function for each input variation, mapping the value of the input variable to the membership degree of the fuzzy set. and are inputs, and the Gaussian function is selected as the membership function. and The maximum value of is 1, and the minimum value is 0. The improvement of the ANFIS prediction model according to the prediction difficulty is located in the fuzzification layer. The second layer is the rule layer, which obtains the excitation intensity of each rule and the fuzzy set of the output variable by performing logical operations on the membership degrees of the input variables. The third layer is the normalization layer, which normalizes the excitation intensity of each rule obtained in the previous layer to obtain its normalized weight. The number of normalized weights is the same as the number of fuzzy rules. The fourth layer is the inference layer, which executes the consequent part of the fuzzy rules and infers the output value from the rules. The fifth layer is the output layer, which performs a weighted average on the output values of each rule to obtain the exact output result.

[0065] It should also be noted that the membership function of the ANFIS network is dynamically adjusted according to the prediction difficulty value D, enabling the network to better adapt to frequency modulation command sequences with different complexities. Through the dynamic adjustment of the membership function and fuzzy rules, the improved ANFIS network can more efficiently process non-linear frequency modulation command sequences and reduce prediction errors. The improved ANFIS network can more accurately predict the value of the frequency modulation command by combining the interpretability of the fuzzy inference system and the learning ability of the adaptive network. By dynamically adjusting the network structure and optimizing the prediction process, the improved ANFIS network can generate prediction results faster, thereby reducing the response time of the energy storage system.

[0066] S3: Predict the value of the frequency modulation command through the improved ANFIS network and output the result.

[0067] Further, predicting the value of the frequency modulation command and outputting the result includes inputting the frequency modulation command sequence into the improved ANFIS network for prediction. The improved ANFIS network dynamically adjusts the network structure according to the prediction difficulty value D and outputs the prediction result; taking the prediction result as the final value for the energy storage system to act in advance.

[0068] It should be noted that the improved ANFIS network can adaptively adjust according to the prediction difficulty value of the frequency modulation command sequence, enhancing the flexibility and applicability of the model; by improving the prediction accuracy and reducing the response time, the improved ANFIS network reduces the dependence on standby power generation equipment, thereby reducing the frequency modulation cost and improving the economic benefits of the power plant; the improved ANFIS network can better capture the local fluctuation characteristics of the frequency modulation command sequence, optimizing the frequency modulation response speed and stability of the power plant.

[0069] Example 2, referring to Figure 3 , which is the second embodiment of the present invention, provides a method for predicting the energy storage frequency modulation command based on the radius method. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0070] First, obtain the frequency modulation commands, frequency modulation sequence 1 and frequency modulation sequence 2 from a certain power plant. Among them, the collection time of frequency modulation sequence 1 is from 0:00 to 15:00 on a certain day of a certain month, and one data point is collected every 1 second; the collection time of frequency modulation sequence 2 is from 0:00 to 15:00 on the 2nd day of a certain month, and one data point is collected every 1 second.

[0071] Secondly, the experiment respectively uses the method of the present invention and the prediction method of ANFIS to predict the frequency modulation sequence. The present invention evaluates the prediction difficulty value of the frequency modulation command sequence by the radius method, adjusts the structure of the improved ANFIS network according to the obtained difficulty value, predicts the value of the frequency modulation command based on the improved ANFIS prediction model and outputs the result, records the experimental data and calculates the prediction error according to the evaluation index, and conducts a comparative analysis with the prediction method of ANFIS. Among them, the experimental data of the prediction error is shown in Table 1 below.

[0072] Table 1 Experimental data of prediction error

[0073] ,

[0074] The following Table 2 shows four evaluation indexes.

[0075] Table 2 Four evaluation indexes

[0076] ,

[0077] Among them, MAE represents the mean absolute prediction error value, SSE represents the sum of squared errors within the group, RMSE represents the root mean square prediction error value, and MAPE represents the mean absolute error.

[0078] It can be seen from the experimental data that through the prediction difficulty evaluation based on the radius method and the improved ANFIS prediction model, the experimental data shows that the accuracy of the prediction results has been significantly improved. Compared with the unimproved ANFIS method, the improved ANFIS network can capture the complexity and nonlinear characteristics of the frequency modulation command sequence more accurately, thereby reducing the prediction error. Specifically, the fitting degree between the prediction result and the measured value is higher, and the error rate is reduced.

[0079] Through the dynamic adjustment of the ANFIS network structure based on the prediction difficulty value, the model can adapt to frequency modulation command sequences with different complexities. Experiments show that the improved model is more stable, reliable, and adaptable when dealing with high-volatility and nonlinear frequency modulation commands.

[0080] Embodiment 3 is the third embodiment of the present invention. What is different from the previous two embodiments is:

[0081] If the said function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical disks, etc., which can store program codes.

[0082] The logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or used in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device.

[0083] More specific examples (a non-exhaustive list) of computer-readable media include the following: electrical connections (electronic devices) having one or more wirings, portable computer disk cartridges (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber devices, and portable compact disc read-only memory (CDROM). Additionally, the computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then stored in a computer memory.

[0084] It should be understood that the various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one or a combination of the following techniques well-known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having suitable combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0085] Example 4, referring to Figure 3 , is the fourth embodiment of the present invention. This embodiment provides a system for predicting energy storage frequency modulation commands based on the radius method, including a frequency modulation evaluation module, a difficulty network module, and a prediction output module.

[0086] The frequency modulation evaluation module is used to evaluate the prediction difficulty value of the frequency modulation command sequence by the radius method; the difficulty network module is used to improve the ANFIS prediction network structure according to the prediction difficulty value; the prediction output module is used to predict the value of the frequency modulation command through the improved ANFIS network and output the result.

[0087] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. A method for predicting energy storage frequency modulation commands based on the radius method, characterized in that: including evaluating the prediction difficulty value of the frequency modulation command sequence by the radius method improving the ANFIS prediction network structure according to the prediction difficulty value predicting the value of the frequency modulation command by the improved ANFIS network and outputting the result The prediction difficulty value includes a prediction difficulty evaluation, expressed as: where D represents the prediction difficulty value, K represents the number of centers that simultaneously satisfy two similarities, and N represents the number of times of frequency modulation command acquisition The evaluating the prediction difficulty value of the frequency modulation command sequence by the radius method includes determining the center position and radius length of adjacent two points in the frequency modulation command sequence in the rectangular coordinate system Let two adjacent points be X i and X i+1 , and the coordinate values of the two points in the rectangular coordinate system are A(i, X i ) and B(i + 1, X i+1 ). Connect A and B, and draw the perpendicular bisector of line segment AB. Then the centers of the two adjacent points are on the straight line l, and the straight line l is expressed as: Let the centers of two adjacent points be o (i) , o (i) has been moving downward on the center line l and forms a circle with the distance from o (i) to X i (X i+1 ) as the radius. By dynamically adjusting the center position, the line segment AB divides the circle into two parts S1 and S2. During the process of forming the circle, for the areas of the two parts S1 and S2, when S1 / S2 < 0.03, the radius length is R(i) and the center is o (i) ; The prediction difficulty value of the frequency modulation command sequence evaluated by the radius method further includes obtaining the frequency modulation command sequence P by the radius method t in the radius sequence [R1, R2, R3,..., R N-1 and the center sequence [o1, o2, o3,..., o N-1 ; conducting similarity analysis on the radii and centers of all adjacent points in the frequency modulation command sequence to find points with similar center coordinates and similar radii Points with similar center coordinates include calculating the distances between two points in the center sequence, and setting the maximum distance as L max , and setting a center o i The distances to other centers are all less than L max / 2, then taking o i as the center and L max / exp(2i) as the radius to draw a circle, and defining the centers that fall within the circle with radius L max / exp(2i) as the centers with similar coordinates, and setting the quantity as S; Radius similarity includes setting the radius sequence [R1, R2, R3, …, R N-1 , where the radius of the circle with the largest radius in the sequence is R max , and the smallest radius is R min , and the average value is avg[R1, R2, R3, …, R N-1 . If the radius R i satisfies the similarity interval, then the radii within the similarity interval satisfy radius similarity, and the number is set to A; similarity interval, expressed as: [avg[R1,R2,R3,…,R N-1 -sin((R min +R max ) / 2),avg[R1,R2,R3,...,R N-1 +sigmoid(R max )] If a center satisfies both center similarity and radius similarity, the number that simultaneously satisfies the two similarities is denoted as K The improvement of the ANFIS prediction network structure according to the prediction difficulty value includes adjusting the membership function of the ANFIS network based on the prediction difficulty value D and Improve and adjust the membership function according to the prediction difficulty The improved membership function is expressed as: Among them, a i , c i respectively represent the width and center of the Gaussian function, represents the value adjusted and improved according to the prediction difficulty membership function and is expressed as: Improve and adjust the membership function according to the prediction difficulty The improved membership function is expressed as: Among them, represents the value adjusted and improved according to the prediction difficulty membership function and is expressed as: and has a maximum value of 1 and a minimum value of 0.

2. The method for predicting energy storage frequency modulation commands based on the radius method according to claim 1, wherein: The frequency modulation command sequence includes setting the frequency modulation command as P t , which is expressed as: P t = [X1, X2, X3,..., X i ,..., X N ​ where X represents the value of the frequency modulation command, and i and N respectively represent different acquisition times The sampling interval of the frequency modulation command is 1 second, and the value X in the rectangular coordinate system i has a coordinate value of (i, X i ).

3. The method for predicting energy storage frequency modulation commands based on the radius method according to claim 2, wherein: The prediction difficulty value further includes that the larger the value of D, the greater the prediction difficulty 4. The method for predicting energy storage frequency modulation commands based on the radius method according to claim 3, characterized in that: The predicting the value of the frequency modulation command and outputting the result includes inputting the frequency modulation command sequence into the improved ANFIS network for prediction. The improved ANFIS network dynamically adjusts the network structure according to the prediction difficulty value D and outputs the prediction result using the prediction result as the final value for the energy storage system to act in advance 5. A system for predicting energy storage frequency modulation commands based on the radius method, applying a method for predicting energy storage frequency modulation commands based on the radius method as described in any one of claims 1 to 4, characterized in that, including: a frequency modulation evaluation module, a difficulty network module, and a prediction output module The frequency modulation evaluation module is used to evaluate the prediction difficulty value of the frequency modulation command sequence by the radius method The difficulty network module is used to improve the ANFIS prediction network structure according to the prediction difficulty value The prediction output module is used to predict the value of the frequency modulation command by the improved ANFIS network and output the result 6. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for predicting energy storage frequency modulation commands based on the radius method according to any one of claims 1 to 4 7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for predicting energy storage frequency modulation commands based on the radius method according to any one of claims 1 to 4

Images