A method and system for predicting frequency regulation commands for energy storage based on command decomposition.

Abstract

This invention discloses a method and system for predicting frequency regulation commands in energy storage systems by command decomposition, belonging to the technical field of frequency regulation command prediction for energy storage systems. The method first obtains the original frequency regulation command sequence and calculates the corresponding trend change angle. Then, based on the trend change angle, the original frequency regulation command sequence is decomposed into a vertical trend sequence and a horizontal trend sequence. Next, the fluctuation of these two sequences is calculated separately. Based on the difference between the horizontal and vertical trend fluctuations, the original frequency regulation command sequence is corrected. Finally, the corrected sequence is input into a GRU network for prediction, thereby obtaining the predicted frequency regulation command sequence. This method improves the accuracy and reliability of prediction by decomposing and correcting the frequency regulation command sequence.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of frequency regulation instruction prediction of energy storage systems, and relates to a kind of instruction decomposition energy storage frequency regulation instruction prediction method and system. BACKGROUND

[0002] With the continuous development of power systems, hybrid energy storage systems gradually become an indispensable part of modern power grids due to their efficient frequency regulation capability and flexible response speed. Hybrid instruction decomposition energy storage frequency regulation instruction prediction technology aims to establish an accurate prediction model by in-depth analysis of historical frequency regulation instructions and related data to predict future frequency regulation instruction requirements, thereby optimizing the control strategy of hybrid energy storage systems and further improving their frequency regulation performance and efficiency.

[0003] However, in practical applications, traditional frequency regulation instruction prediction methods face many challenges. The traditional approach is to directly input the original frequency sequence into a neural network such as GRU (Gated Recurrent Unit) for prediction. However, the original frequency sequence often exhibits high nonlinearity and irregularity, which results in low accuracy of direct prediction results and large prediction errors.

[0004] To overcome this difficulty, researchers have attempted to preprocess the original frequency sequence to extract more regular features. Among them, VMD (Variational Mode Decomposition) and CEEMD (Complementary Ensemble Empirical Mode Decomposition) are two commonly used preprocessing methods. VMD decomposes the original signal into a series of modal components with specific frequencies, helping to reveal the potential regularity hidden in complex signals. CEEMD further improves the stability and accuracy of decomposition by adding noise and multiple decompositions.

[0005] Although VMD and CEEMD have achieved remarkable results in signal processing, there are still a series of problems when combining them for decomposition and prediction of original frequency sequences. First, the computational complexity of VMD and CEEMD is high, and the decomposition process is time-consuming, especially for long time series data, with poor real-time performance. In addition, the parameter tuning process of both is complex, and improper parameter selection can lead to over-decomposition or under-decomposition, affecting the accuracy of the prediction results.

[0006] Secondly, VMD and CEEMD may produce modal redundancy and over-decomposition risk in the decomposition process. If the modal number (K) of VMD is set too large, noise modal without physical meaning may be decomposed; and if it is set too small, key features cannot be captured. CEEMD may produce redundant modal due to noise residue or over-decomposition, resulting in too many sub-sequences and increasing the complexity of the subsequent prediction model.

[0007] In addition, modal aliasing and frequency confusion are also problems that VMD and CEEMD need to solve in the decomposition of frequency modulation sequences. Although CEEMD reduces the modal aliasing problem of traditional EMD (Empirical Mode Decomposition), it may still overlap in the mutation or high-frequency noise area of the original frequency modulation sequence. VMD assumes that the modal is a narrowband signal, and the actual frequency modulation sequence may contain wideband components, so the frequency boundary between the modes after decomposition is ambiguous.

[0008] In terms of prediction model integration, each sub-sequence after VMD and CEEMD decomposition needs to be modeled or predicted separately, and the number of models increases exponentially, and the integration strategy may introduce errors. At the same time, the sub-sequences after decomposition need to be time-aligned, and if there is a phase shift in the decomposition process, the prediction result may be distorted.

[0009] Finally, VMD and CEEMD may also cause information loss and reconstruction error in the decomposition process. When the sub-modes after decomposition are reconstructed into the original signal, decomposition errors or improper parameters will cause the reconstructed signal to deviate from the true value, affecting the quality of the prediction input. In addition, random fluctuations in the original frequency modulation sequence may be decomposed into high-frequency modes, and if they are simply ignored or processed, key transient information may be lost.

[0010] In summary, the traditional VMD and CEEMD decomposition methods have many shortcomings when dealing with original frequency modulation sequences, such as high computational complexity, modal redundancy and over-decomposition risk, modal aliasing and frequency confusion, difficulty in integrating prediction models, information loss and reconstruction error, and insufficient handling of mutations and non-stationarity. SUMMARY

[0011] The purpose of the present application is to solve the technical problems in the prior art that the VMD and CEEMD decomposition methods have high computational complexity when processing original frequency modulation sequences, and the large data fluctuations cause low prediction accuracy when used for prediction, and to provide an instruction decomposition energy frequency modulation instruction prediction method and system.

[0012] To achieve the above-mentioned purpose, the following technical solutions are adopted in the present application:

[0013] The present application discloses an instruction decomposition energy frequency modulation instruction prediction method in the first aspect, comprising the following steps:

[0014] obtaining an original frequency modulation instruction sequence;

[0015] The corresponding trend change angle is calculated based on the original frequency modulation command sequence; based on the trend change angle, the original frequency modulation command sequence is decomposed to obtain the vertical trend sequence and the horizontal trend sequence;

[0016] Calculate horizontal trend fluctuations based on horizontal trend sequences; calculate vertical trend fluctuations based on vertical trend sequences.

[0017] The original frequency modulation command sequence is modified according to the different horizontal and vertical trend fluctuations.

[0018] The corrected original frequency modulation command sequence is input into the GRU network for prediction to obtain the predicted frequency modulation command sequence.

[0019] Furthermore, the calculation of the corresponding trend change angle based on the original frequency modulation command sequence specifically involves:

[0020]

[0021]

[0022] in, The first of the original frequency modulation command sequence Position to number The angle of trend change in location; It is the arctangent function; It is the Euler number; The first of the original frequency modulation command sequence The value of the position; The first of the original frequency modulation command sequence The value of the position; The first of the original frequency modulation command sequence The value of the position; The length of the original frequency modulation command sequence; Represents absolute value; For the original frequency modulation command sequence, the first... The value of the position; For the original frequency modulation command sequence, the first... The value of the position; For the original frequency modulation command sequence, the first... The value of the position; The first of the original frequency modulation command sequence Position to number Angle of trend change in location.

[0023] Furthermore, based on the trend change angle, the original frequency modulation command sequence is decomposed to obtain a vertical trend sequence and a horizontal trend sequence, specifically:

[0024] The original frequency modulation command sequence number Position to number the vector from the original frequency modulation instruction sequence first position as the origin clockwise by a trend change angle to obtain a longitudinal trend;

[0025] clockwise by 90 degrees based on the longitudinal trend to obtain a transverse trend;

[0026] According to the above method, the original frequency modulation instruction sequence is decomposed to obtain a longitudinal trend sequence and a transverse trend sequence; the value of the first position of the longitudinal trend sequence is described as:

[0027]

[0028] wherein, is a vector from the original frequency modulation instruction sequence first position to the first position; is a trend change angle from the original frequency modulation instruction sequence first position to the first position; is a longitudinal trend from the original frequency modulation instruction sequence first position to the first position; cos() represents a cosine function;

[0029] the value of the first position of the transverse trend sequence is described as:

[0030]

[0031] is a transverse trend from the original frequency modulation instruction sequence first position to the first position; sin() represents a sine function.

[0032] Further, based on the longitudinal trend sequence, a longitudinal trend fluctuation is calculated, specifically:

[0033] the longitudinal trend in the longitudinal trend sequence with a counterclockwise angle with the horizontal direction less than 180 degrees is multiplied by to obtain a new longitudinal trend sequence;

[0034] the sum of each item of the new longitudinal trend sequence is obtained to obtain a longitudinal trend fluctuation .

[0035] Further, based on the transverse trend sequence, a transverse trend fluctuation is calculated, specifically:

[0036] based on the transverse trend sequence, the median and the average of the transverse trend sequence are calculated;

[0037] The horizontal trend fluctuation is calculated based on the median and mean of the horizontal trend sequence; the horizontal trend fluctuation is:

[0038]

[0039] in, `min()` is the maximum value function; `min()` is the minimum value function. This indicates a horizontal trend fluctuation.

[0040] Furthermore, based on the different horizontal and vertical trend fluctuations, the original frequency modulation command sequence is modified, specifically as follows:

[0041] when and If necessary, the original frequency modulation command sequence is modified as follows: ;

[0042] when and If necessary, the original frequency modulation command sequence is modified as follows: ;

[0043] when and If necessary, the original frequency modulation command sequence is modified as follows: ;

[0044] in, The corrected original frequency modulation command sequence is the first... The value of the position; The first of the original frequency modulation command sequence The value of the position; It is the Euler number; The first of the original frequency modulation command sequence Position to number A vector of position; The first of the original frequency modulation command sequence Position to number The angle of change in position; cos() is the cosine function; sin() is the sine function; tanh() is the tangent function; This is the activation function.

[0045] A second aspect of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements a method for predicting energy storage frequency modulation instructions by decomposing the instructions.

[0046] The third aspect of the present application provides a computer readable storage medium, the computer readable storage medium stores a computer program, the computer program is executed by a processor to realize the energy storage frequency modulation instruction prediction method of instruction decomposition.

[0047] The fourth aspect of the present application provides a computer program product, the computer program product comprises computer instructions, the computer instructions instruct the computer to execute the energy storage frequency modulation instruction prediction method of instruction decomposition.

[0048] The fifth aspect of the present application provides an energy storage frequency modulation instruction prediction system of instruction decomposition, comprising:

[0049] The data acquisition module acquires the original frequency modulation instruction sequence;

[0050] The decomposition module calculates the corresponding trend change angle based on the original frequency modulation instruction sequence; based on the trend change angle, the original frequency modulation instruction sequence is decomposed to obtain the longitudinal trend sequence and the transverse trend sequence;

[0051] The fluctuation estimation module calculates the transverse trend fluctuation based on the transverse trend sequence; the longitudinal trend fluctuation is calculated based on the longitudinal trend sequence;

[0052] The correction module corrects the original frequency modulation instruction sequence according to the difference between the transverse trend fluctuation and the longitudinal trend fluctuation;

[0053] The prediction module inputs the corrected original frequency modulation instruction sequence into the GRU network for prediction to obtain the predicted frequency modulation instruction sequence.

[0054] Compared with the prior art, the present application has the following beneficial effects:

[0055] The present application discloses an energy storage frequency modulation instruction prediction method of instruction decomposition, which first decomposes the original frequency modulation instruction sequence into a longitudinal trend sequence and a transverse trend sequence by calculating a trend change angle. This decomposition method helps to capture different change trends in the frequency modulation instruction more carefully, thereby improving the accuracy of prediction. By separately calculating the transverse trend fluctuation and the longitudinal trend fluctuation, this method can more comprehensively understand the fluctuation characteristics of the frequency modulation instruction, providing a more reliable data basis for subsequent correction and prediction. By correcting the original frequency modulation instruction sequence, this method can eliminate or weaken the noise and outliers in the original data, making the corrected frequency modulation instruction sequence smoother and more stable. The corrected frequency modulation instruction sequence is input into the GRU (Gated Recurrent Unit) network for prediction. Since the GRU network has excellent performance in processing time series data, it can further improve the accuracy and real-time performance of the prediction, thereby optimizing the response speed and stability of the energy storage system.

[0056] Further, by rotating each adjacent vector of the original frequency modulation instruction sequence by a trend change angle clockwise with the origin as the reference, the method can accurately separate the longitudinal trend in the frequency modulation instruction, i.e., the main direction of the change of the frequency modulation instruction over time. At the same time, the transverse trend is obtained by further rotation, capturing the subtle fluctuations and potential changes of the frequency modulation instruction. The separation of the longitudinal trend and the transverse trend makes the subsequent fluctuation calculation more accurate, and the two trends represent the changes in frequency and amplitude in the frequency modulation instruction sequence. This refined decomposition helps the prediction model to more accurately capture the change rule of the frequency modulation instruction, thereby improving the accuracy of the prediction. BRIEF DESCRIPTION OF DRAWINGS

[0057] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0058] Figure 1 Block diagram of the energy storage frequency modulation instruction prediction method for instruction decomposition;

[0059] Figure 2 Block diagram of the energy storage frequency modulation instruction prediction system for instruction decomposition.

[0060] Among them: 201-data acquisition module; 202-decomposition module; 203-fluctuation estimation module; 204-correction module; 205-prediction module. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all. The components of the embodiments of the present application described and indicated in the drawings here can be arranged and designed in various different configurations.

[0062] Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0063] It should be noted that: similar numbers and letters represent similar items in the following drawings, therefore, once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.

[0064] The application will be further described in detail below with reference to the accompanying drawings:

[0065] Referring to Figure 1 The application discloses a kind of instruction decomposition energy storage frequency modulation instruction prediction method, comprising the following steps:

[0066] S1, obtain original frequency modulation instruction sequence;From the historical database of grid dispatching system or energy storage system, extract the frequency modulation instruction data in a period of time (such as past a week, a month), form original frequency modulation instruction sequence.These data are usually presented in the form of time series, record the frequency modulation demand of each time point.

[0067] S2, based on original frequency modulation instruction sequence, corresponding trend change angle is calculated;Based on trend change angle, original frequency modulation instruction sequence is decomposed, and longitudinal trend sequence and transverse trend sequence are obtained.

[0068] S3, based on transverse trend sequence, transverse trend fluctuation is calculated;Based on longitudinal trend sequence, longitudinal trend fluctuation is calculated;Original frequency modulation instruction sequence is decomposed into longitudinal trend sequence and transverse trend sequence using trend change angle.Longitudinal trend sequence reflects the long-term growth or reduction trend of frequency modulation instruction with time, and transverse trend sequence captures the fluctuation change in short term.This decomposition helps to more clearly understand the composition of frequency modulation demand, facilitating subsequent analysis.

[0069] S4, according to the difference of transverse trend fluctuation and longitudinal trend fluctuation, the original frequency modulation instruction sequence is modified;

[0070] S5, the original frequency modulation instruction sequence after modification is input into GRU network for prediction, and the predicted frequency modulation instruction sequence is obtained.

[0071] One embodiment of the application discloses a kind of instruction decomposition energy storage frequency modulation instruction prediction method, comprising the following steps:

[0072] S1, obtain original frequency modulation instruction sequence;From the historical database of grid dispatching system or energy storage system, extract the frequency modulation instruction data in a period of time (such as past a week, a month), form original frequency modulation instruction sequence.These data are usually presented in the form of time series, record the frequency modulation demand of each time point.

[0073] S2, based on original frequency modulation instruction sequence, corresponding trend change angle is calculated;Based on trend change angle, original frequency modulation instruction sequence is decomposed, and longitudinal trend sequence and transverse trend sequence are obtained;Based on original frequency modulation instruction sequence, corresponding trend change angle is calculated, specifically:

[0074]

[0075]

[0076] wherein, is the value of the original FM command sequence at the position; is the trend change angle from the is an inverse trigonometric function; is Euler's number; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the length of the original FM command sequence; denotes the absolute value; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the trend change angle from the position to the position.

[0077] rotating the vector from the position to the position of the original FM command sequence by a trend change angle clockwise with the position of the original FM command sequence as the origin, a longitudinal trend is obtained;

[0078] rotating the longitudinal trend by 90 degrees clockwise, a transverse trend is obtained;

[0079] The original FM command sequence is decomposed according to the above method to obtain a longitudinal trend sequence and a transverse trend sequence; the value of the position in the longitudinal trend sequence is described as:

[0080]

[0081] wherein, is the vector from the position to the position of the original FM command sequence; is the trend change angle from the position to the position of the original FM command sequence; is the vector from the position to the Vertical trend of location; Represents the cosine function;

[0082] The horizontal trend sequence number The value of the position is described as follows , indicating the first of the original frequency modulation command sequence Position to number The horizontal trend of the position; sin() represents the sine function.

[0083] S3 calculates the horizontal trend fluctuation based on the horizontal trend sequence and the vertical trend fluctuation based on the vertical trend sequence. Using the trend change angle, the original frequency modulation (FM) command sequence is decomposed into a vertical trend sequence and a horizontal trend sequence. The vertical trend sequence reflects the long-term growth or decrease trend of FM commands over time, while the horizontal trend sequence captures short-term fluctuations. This decomposition helps to more clearly understand the composition of FM demand and facilitates subsequent analysis.

[0084] Multiply all vertical trends in the vertical trend sequence that have a counterclockwise angle of less than 180 degrees with the horizontal direction by [the specified factor]. This yields a new vertical trend sequence;

[0085] Sum each term of the new longitudinal trend sequence to obtain the longitudinal trend fluctuation. .

[0086] Calculate the median of the horizontal trend series. and mean ;

[0087] The horizontal trend fluctuation is calculated based on the median and mean of the horizontal trend sequence; the horizontal trend fluctuation is:

[0088]

[0089] in, `min()` is the maximum value function; `min()` is the minimum value function. This indicates a horizontal trend fluctuation.

[0090] S4, based on the different horizontal and vertical trend fluctuations, the original frequency modulation command sequence is modified;

[0091] when and If necessary, the original frequency modulation command sequence is modified as follows: ;

[0092] when and If necessary, the original frequency modulation command sequence is modified as follows: ;

[0093] when and If t is greater than or equal to 0.5, the original frequency modulation instruction sequence is corrected, specifically: ;

[0094] wherein, is the value of the corrected original frequency modulation instruction sequence at the i th position; is the value of the original frequency modulation instruction sequence at the i th position; is Euler's number; is the vector from the i th position to the j th position of the original frequency modulation instruction sequence; is the trend change angle from the i th position to the j th position of the original frequency modulation instruction sequence; cos() is a cosine function; sin() is a sine function; tanh() is a tangent function; is an activation function. No correction is made to . S5, input the corrected original frequency modulation instruction sequence into the GRU network for prediction to obtain a predicted frequency modulation instruction sequence. It should be noted that, in actual implementation, all parameters are processed by Min-Max Scaling to enable operation. In order to further verify the advantages of the present application, the frequency modulation sequence is predicted by the method of the present application and the GRU prediction method, respectively. In this embodiment, the frequency modulation data of an actual power plant is used for comparison test, and the method of the present application is compared with the commonly used GRU (Gated Recurrent Unit) prediction method. The frequency modulation instruction data of this embodiment is from a power plant in Gansu Province.

[0095] Frequency modulation sequence 1: the data points are collected every 1 second from 4:00 to 18:00 on a certain day in January 2021.

[0096] Frequency modulation sequence 2: the data points are collected every 1 second from 0:00 to 19:00 on a certain day in March 2022.

[0097] The results are as follows:

[0098]

[0099]

[0100]

[0101]

[0102] ​​​​​​From the experimental results, it can be seen that in the prediction of frequency modulation sequence 1 and frequency modulation sequence 2, the method of the present application is significantly better than the GRU method in four evaluation indexes. Specifically, the indexes such as MAE (Mean Absolute Error), SSE (Sum of Squared Errors), RMSE (Root Mean Square Error) and MAPE (Mean Absolute Percentage Error) are greatly reduced, which shows that the prediction model of the present application has higher prediction accuracy and better generalization ability.

[0103] Referring to Figure 2 In an embodiment of the present application, an instruction decomposition energy storage frequency modulation instruction prediction system is provided, comprising:

[0104] A data acquisition module 201 acquires an original frequency modulation instruction sequence;

[0105] A decomposition module 202 calculates a corresponding trend change angle based on the original frequency modulation instruction sequence; based on the trend change angle, the original frequency modulation instruction sequence is decomposed to obtain a longitudinal trend sequence and a transverse trend sequence;

[0106] A fluctuation estimation module 203 calculates a transverse trend fluctuation based on the transverse trend sequence; and calculates a longitudinal trend fluctuation based on the longitudinal trend sequence;

[0107] A correction module 204 corrects the original frequency modulation instruction sequence according to the difference between the transverse trend fluctuation and the longitudinal trend fluctuation;

[0108] A prediction module 205 inputs the corrected original frequency modulation instruction sequence into a GRU network for prediction to obtain a predicted frequency modulation instruction sequence.

[0109] In an embodiment of the present application, an electronic device is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor; the device is particularly designed to execute the instruction decomposition energy storage frequency modulation instruction prediction method. The device can be any device with computing capability, such as a personal computer, a workstation, a server, an embedded system, a smart terminal, etc. When the electronic device is started, the processor will load and execute the computer program of the instruction decomposition energy storage frequency modulation instruction prediction method from the memory.

[0110] The application further discloses a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to realize the energy storage frequency modulation instruction prediction method of instruction decomposition; the storage medium of the application is a non-volatile or readable and writable data storage device, which is designed to store and retain computer program instructions. The storage medium can be, but is not limited to, a hard disk drive (HDD, Hard-Disk Drive), a solid state disk (SSD, Solid State Disk), a flash drive (USB, Universal Serial Bus), an optical disc (CD, Compact Disc) (such as a read-only optical disc (CD-ROM, Compact Disc Read-Only Memory), a digital video disc (DVD, Digital Video Disc)), a virtual storage space in a cloud storage service, or any other medium capable of storing and allowing a computer to access and execute program code.

[0111] An embodiment of the application provides a computer program product, which contains computer instructions for guiding a computer to execute the energy storage frequency modulation instruction prediction method of instruction decomposition. The computer program product can exist in various forms, including but not limited to executable files, source code files, script files, installation packages, compressed packages, etc. The product can be stored on a physical medium, such as a hard disk, an optical disc, etc., or can be transmitted and distributed through a network.

[0112] The above is only a preferred embodiment of the application and is not used to limit the application. The application can have various changes and modifications for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. A method for predicting energy storage frequency modulation commands through command decomposition, characterized in that, The method comprises the following steps: obtaining an original frequency modulation instruction sequence; calculating a corresponding trend change angle based on the original frequency modulation instruction sequence; based on the trend change angle, decomposing the original frequency modulation instruction sequence to obtain a longitudinal trend sequence and a transverse trend sequence; calculating a transverse trend fluctuation based on the transverse trend sequence; calculating a longitudinal trend fluctuation based on the longitudinal trend sequence; modifying the original frequency modulation instruction sequence according to the difference between the transverse trend fluctuation and the longitudinal trend fluctuation; inputting the modified original frequency modulation instruction sequence into a GRU network for prediction to obtain a predicted frequency modulation instruction sequence; The method comprises the following steps: wherein, is the value of the original FM command sequence at the position; is the trend change angle from the is the arctangent function; is Euler's number; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the length of the original FM command sequence; denotes the absolute value; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is the trend change angle from the The method comprises the following steps: The original frequency modulation command sequence number Position to number The position vector is the first in the original frequency modulation command sequence. The vertical trend is obtained by rotating the position clockwise from the origin by a trend change angle. The method comprises the following steps: The original frequency modulation instruction sequence is decomposed according to the method to obtain a longitudinal trend sequence and a transverse trend sequence; the value of the position in the longitudinal trend sequence is described as: The value of the position is described as: wherein is the original frequency modulation command sequence at the position to the position; is the original frequency modulation command sequence at the position to the position; is the original frequency modulation command sequence at the position to the position; cos() denotes the cosine function; The lateral trend sequence is described as: The value of the position is described as: for the original FM command sequence position to the position of the lateral trend; sin() denotes the sine function; The method comprises the following steps: multiplying each of the longitudinal trends in the longitudinal trend sequence that has an anticlockwise angle with the horizontal direction less than 180 degrees by , to obtain a new longitudinal trend sequence; Summing each item of the new longitudinal trend series gives the longitudinal trend fluctuation ; The method comprises the following steps: based on the lateral trend sequence, calculating a median of the lateral trend sequence and the average ; The method comprises the following steps: wherein is a maximum function; min() is a minimum function; denotes the lateral trend fluctuation.

2. The instruction-decomposed EDF instruction prediction method of claim 1, wherein, The method comprises the following steps: When and , then the original frequency modulation instruction sequence is corrected, specifically: ; When and then the original frequency modulation command sequence is modified, in particular: ; When and then the original frequency modulation command sequence is modified, in particular: ; wherein, is the value of the original FM command sequence at the position; is the value of the original FM command sequence at the position; is Euler's number; is the vector from the position to the position of the original FM command sequence; is the trend change angle from the position to the position of the original FM command sequence; cos() is the cosine function; sin() is the sine function; tanh() is the hyperbolic tangent function; is the activation function; is the longitudinal trend fluctuation.

3. An electronic device, comprising: The method comprises the following steps:

4. A computer-readable storage medium, characterized in that, The method comprises the following steps:

5. A computer program product comprising computer instructions, characterized in that, The method comprises the following steps:

6. An instruction-decomposed energy storage frequency modulation instruction prediction system based on the instruction-decomposed energy storage frequency modulation instruction prediction method of claim 1, characterized in that, The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method comprises the following steps: The method

Images