Energy storage frequency modulation instruction prediction method and system of substitution method

Abstract

The energy storage frequency modulation instruction prediction method comprises the following steps: acquiring an original frequency modulation instruction signal sequence; finding out a strong fluctuation point based on the original frequency modulation instruction signal sequence; replacing the strong fluctuation point signal to generate a replacement signal sequence; using the original frequency modulation instruction signal sequence to replace the signal sequence for prediction to obtain a predicted frequency modulation instruction sequence; performing power regulation and control on different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence; the prediction result of the frequency modulation instruction is more accurate, the size of the frequency modulation instruction is predicted in advance for energy storage regulation and control, the super capacitor / battery can act in advance, and the response accuracy and the frequency modulation benefit are improved.

Description

Technical Field

[0001] The present disclosure relates to the technical field of power grid frequency regulation, and in particular to a method and system for predicting energy storage frequency regulation instructions by substitution. Background Art

[0002] At present, in major regional power grids in my country, large hydropower and thermal power units (coal-fired units / gas-fired units) are mainly used as power grid frequency regulation power sources. The frequency regulation power output is adjusted to respond to system frequency changes. However, hydropower and thermal power units have certain limitations in frequency regulation. In recent years, the increase in the proportion of new energy power generation such as wind power and solar power has led to a large number of thermal power units bearing heavy AGC regulation tasks for a long time, resulting in a series of negative effects such as increased coal consumption for power generation and serious equipment wear. The existing power frequency regulation resources can no longer meet the demand for renewable energy access to the grid. The hybrid energy storage system with supercapacitors participates in the coordinated control of primary frequency regulation. By dividing the typical working conditions, the frequency deviation is converted to power according to the optimization decision module, and the charge / discharge control of the supercapacitor energy storage system is realized through the PCS system control module, and finally assists the traditional generator set to complete the primary frequency regulation process. Within the rated power range, the specified power output can be completed within 1s with an accuracy of more than 99%. Its comprehensive response capability fully meets the power conversion requirements within the time scale of AGC frequency regulation, which greatly exceeds the regulation capacity of conventional thermal power plants. The more urgent the system's regulation needs, the more obvious the advantages of this technology.

[0003] The hybrid energy storage system based on the above-mentioned supercapacitor energy storage usually uses VMD (Variational mode decomposition) algorithm to realize the distribution of frequency modulation signals. The traditional frequency modulation method of hybrid energy storage (supercapacitor + lithium battery) to assist thermal power units is to pass the difference between the frequency modulation instruction and the thermal power unit to the hybrid energy storage, and use VMD decomposition technology to decompose the original signal into K subsequences IMF1 to IMFK, of which the low-frequency part is borne by the battery and the high-frequency part is borne by the supercapacitor. VMD is an adaptive, completely non-recursive modal variation and signal processing method. This technology has the advantage of being able to determine the number of modal decompositions. Its adaptability is reflected in determining the number of modal decompositions of a given sequence according to actual conditions. In the subsequent search and solution process, the optimal center frequency and limited bandwidth of each mode can be adaptively matched, and the effective separation of the intrinsic modal component (IMF) and the frequency domain division of the signal can be achieved, thereby obtaining the effective decomposition components of the given signal, and finally obtaining the optimal solution to the variational problem. The decomposition obtains a relatively stable subsequence containing multiple different frequency scales, which is suitable for non-stationary sequences.

[0004] The traditional hybrid energy storage (supercapacitor + lithium battery) assists the frequency regulation of thermal power units by transmitting the difference between the frequency regulation command and the thermal power unit to the hybrid energy storage, where the low-frequency part is borne by the battery and the high-frequency part is borne by the supercapacitor. However, the transmission of the signal (i.e., the transmission of the frequency regulation command to the supercapacitor / lithium battery) takes time, and the supercapacitor or lithium battery itself also takes time to respond, which will cause a certain response time difference, further affecting the accuracy of the system response and the frequency regulation benefits of the power plant. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes a prediction method to predict the size of the frequency modulation instruction in advance, so that the supercapacitor / battery can be activated in advance to improve the response accuracy and frequency modulation benefit.

[0006] A method for predicting energy storage frequency modulation instructions by substitution method, characterized by comprising the steps of: S101, obtaining an original frequency modulation instruction signal sequence; S102, finding strong fluctuation points based on the original frequency modulation instruction signal sequence; S103, replacing the strong fluctuation point signal to generate a replacement signal sequence; S104, using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction to obtain a predicted frequency modulation instruction sequence; S105. Perform power control of different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence.

[0007] Further, In step S101, Dividing the original frequency modulation instruction signal sequence into signal sequences and forming groups; The frequency modulation command signal sequence is Pt=[X1,X2,X3,.X i ..,X N ], the frequency modulation command signal sequence is divided into two parts according to the length ratio, the front part is used as the input group sequence, and the back part is used as the test group sequence.

[0008] Further, If the length ratio is 0.9, then the integer sequence [X1, X2, X3....., X 0.9N ] as the input group, followed by an integer sequence of 10% length [X 0.9N+1 ,X 0.9N+2 ,X 0.9N+3 ,...,X N ] as the test group; construct the unknown group sequence, with [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ] as the unknown group.

[0009] Further, The above S102, finding the strong fluctuation point based on the original frequency modulation instruction signal sequence, includes: Calculating the change slopes and change angle values ​​of adjacent signals, and determining that a number of signal values ​​whose change angle values ​​are greater than a first angle threshold form a first change signal sequence; After the first change signal sequence is subjected to exponential factorial operation, the change slope and change angle value of the adjacent signals are calculated in the same manner, and a number of signal values ​​whose change angle values ​​are greater than the second angle threshold are determined to form a strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ].

[0010] Further, The above S103, replacing the strong fluctuation point signal to generate a replacement signal sequence, includes: Performing a fitting operation on the strongly changing signal sequence to determine a fitting straight line X=Kt+B; Using parameters K and B, we calculate the strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ], each replacement value is inserted into the original frequency modulation instruction signal sequence to obtain the frequency modulation instruction replacement signal sequence Pt 、 .

[0011] Further, The replacement value calculation formula is: .

[0012] Further, The above S104, using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction to obtain a predicted frequency modulation instruction sequence, comprises: Replace the signal sequence Pt with the frequency modulation instruction 、 Input GRU for prediction and get the first predicted value of the unknown group [X N+1 、 ,X N+2 、 ,X N+3 、 ,...,X N+0.1N 、 ], and the first error W1 of the experimental group was also obtained; The difference sequence Pt-Pt of the original frequency modulation instruction signal sequence and the frequency modulation instruction replacement signal sequence 、Input GRU for prediction and get the second predicted value of the unknown group [X N+1 、、 ,X N+2 、、 ,X N+3 、、 ,...,X N+0.1N 、、 ], the second error W2 of the experimental group; The first prediction value, the second prediction value, the first error, and the second error are integrated to calculate the final prediction result.

[0013] Further, The fusion algorithm of the final prediction result is:

[0014] Get the predicted frequency modulation instruction sequence [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ].

[0015] Further, For strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] to perform straight line fitting, First, randomly generate a fitting straight line X = Kt + B, and calculate the strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] to this straight line is [d1,d2,d3,..,dr], and the sum of [d1,d2,d3,..,dr] is calculated; The loop is executed to randomly generate M fitting straight lines X=Kt+B, where K and B are both randomly generated; Determine the deviation and the smallest fitting straight line among the M fitting straight lines, and use the fitting straight line X=Kt+B as the selected fitting straight line.

[0016] Further, During the prediction process, the input group sequence is used as input to predict the predicted values ​​of the test group and the unknown group, and the average MAPE of the absolute error is calculated based on the actual value and the predicted value of the test group to generate the first error W1 and the second error W2.

[0017] A replacement method energy storage frequency modulation instruction prediction system, characterized by comprising: The command acquisition module acquires the original frequency modulation command signal sequence; The search module finds the strong fluctuation points based on the original frequency modulation command signal sequence; The substitution module replaces the strong fluctuation point signal to generate a substitution signal sequence; The prediction module uses the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction and obtain the predicted frequency modulation instruction sequence; The power control module performs power control of different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence.

[0018] The beneficial effects of the present invention are as follows: a method and system for predicting energy storage frequency modulation instructions by substitution method are provided, which can respond in time when the frequency of the power plant power grid fluctuates, by finding strong fluctuation points, generating alternative signal sequences, and obtaining predicted frequency modulation instruction sequences, thereby making the prediction results of the frequency modulation instructions more accurate, and predicting the size of the frequency modulation instructions in advance for energy storage regulation, which can enable the supercapacitor / battery to act in advance, thereby improving the response accuracy and frequency modulation benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 A diagram of the method steps.

[0020] Figure 2 , Figure 3 This is an illustration of the algorithm for searching for strong fluctuation points. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0022] It should be understood that the step numbers used in this document are only for convenience of description and are not intended to limit the order in which the steps are executed.

[0023] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0024] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0025] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0026] The embodiment of the present disclosure provides a method for predicting energy storage frequency modulation instructions by substitution method, such as Figure 1 As shown, including: S101, obtaining an original frequency modulation instruction signal sequence; S102, finding strong fluctuation points based on the original frequency modulation instruction signal sequence; S103, replacing the strong fluctuation point signal to generate a replacement signal sequence; S104, using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction to obtain a predicted frequency modulation instruction sequence; S105. Perform power control of different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence.

[0027] Wherein, step S101, obtaining the original frequency modulation instruction signal sequence, includes: The original frequency modulation instruction signal sequence is divided and constructed into signal sequences.

[0028] The frequency modulation command signal sequence is Pt=[X1,X2,X3,.X i ..,X N ] The frequency modulation command signal sequence is divided into two parts according to the length ratio, the front part is used as the input group sequence, and the back part is used as the test group sequence.

[0029] If the length ratio value is 0.9, then the integer sequence [X1, X2, X3....., X 0.9N ] as the input group, followed by an integer sequence of 10% length [X 0.9N+1 ,X 0.9N+2 ,X 0.9N+3 ,...,X N ] as the experimental group.

[0030] At the same time, the unknown group sequence is constructed, with [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ] is taken as the unknown group, and the unknown group has the same sequence length as the test group.

[0031] In the prediction process, the input group sequence is used as the input of the neural network to predict the values ​​of the test group and the unknown group. Finally, the value of the unknown group sequence is used to control the power in the hybrid energy storage system.

[0032] In another embodiment disclosed in the present application, Figure 2 As shown, The volatility of the original frequency modulation command signal sequence is very obvious, so signal preprocessing is required. First, find the strong fluctuation points. For the signal sequence, if the angle of the line connecting the points changes greatly, it means that the changes of the two adjacent points are large. Large changes indicate poor regularity and large signal volatility.

[0033] The above S102, finding the strong fluctuation point based on the original frequency modulation instruction signal sequence, includes: Step 1: Calculate the change slopes and change angle values ​​of adjacent signals, and determine that a number of signal values ​​whose change angle values ​​are greater than a first angle threshold form a first change signal sequence.

[0034] The original frequency modulation command signal sequence is Pt=[X1,X2,X3,.X i ..,X N ]. The original frequency modulation command signal sequence Pt is a function of time t, such as Figure 2 As shown in the coordinate system, the horizontal axis represents time t, and the vertical axis represents the signal value X1, X2, X3,.X i ..,X N .

[0035] The change slope of the line segment formed by two adjacent signal values ​​is calculated by the horizontal coordinate and the vertical coordinate of each signal value, and the change angle value formed by two adjacent line segments formed by three adjacent signals is further calculated to obtain the first angle sequence .

[0036] Each angle value in the first angle sequence is compared with the first angle threshold, and a number of signal values ​​corresponding to the values ​​greater than the first angle threshold are obtained to form a first change signal sequence.

[0037] Set the first angle threshold to 50°, select a number of signal values ​​corresponding to angles greater than 50°, and set the first change signal sequence formed by the number of signal values ​​to be [X 1d ,X 2d ,X 3d ,.X id ..,X dd ], where there are d signal values, and their corresponding angles are greater than 50°.

[0038] Step 2: After performing an exponential factorial operation on the first change signal sequence, calculate the change slope and change angle value of adjacent signals, and determine that a number of signal values ​​whose change angle value is greater than a second angle threshold form a strong change signal sequence.

[0039] First, each value in the first change signal sequence is subjected to exponential factorial operation to form the following exponential factorial sequence: [exp(X 1d ) / 1d!,exp(X 2d) / 2d!,exp(X 3d ) / 3d!,.exp(X id ) / id!..,exp(X dd ) / dd!].

[0040] Further Figure 3 As shown, the horizontal axis represents time t, and the vertical axis represents the exponential factorial signal value. The change slope of the line segment formed by two adjacent signal values ​​is further calculated, and the change angle value formed by three adjacent signals is further calculated to obtain the second angle sequence.

[0041] Each angle value in the second angle sequence is compared with the second angle threshold, and a number of signal values ​​corresponding to the angle values ​​greater than the second angle threshold are obtained to form a strong change signal sequence.

[0042] The second angle threshold is set to 30°, and a number of signal values ​​corresponding to angles greater than 30° are selected. The strong change signal sequence formed by the signal values ​​is set to [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ], there are r signal values, and the corresponding angle is greater than 30°.

[0043] The r signal values ​​in the strongly changing signal sequence are taken as the points with the strongest volatility in the original sequence.

[0044] The above S103, replacing the strong fluctuation point signal to generate a replacement signal sequence, includes: Step 1: performing a fitting operation on the strongly changing signal sequence.

[0045] For strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] to perform straight line fitting.

[0046] First, randomly generate a fitting straight line X = Kt + B, and calculate the strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] to this straight line is [d1,d2,d3,..,dr], and the sum of [d1,d2,d3,..,dr] is calculated; Loop and randomly generate M fitting lines X=Kt+B, randomly generate K and B, and repeat M times (M is not less than 100); Determine the deviation and the smallest fitting straight line among the M fitting straight lines, and use the fitting straight line X=Kt+B as the selected fitting straight line.

[0047] Step 2: Replace the signal value in the strongly changing signal sequence.

[0048] Using the parameters K and B in the selected fitting line X=Kt+B, the strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr For points with strong volatility, the coefficients of the fitting line are used to fit the points with strong volatility, thereby enhancing the regularity of these points.

[0049] The replacement value calculation formula is:

[0050] For strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] According to the replacement value calculation formula, a strong change replacement signal sequence is calculated, and the strong change replacement signal sequence is inserted into the original frequency modulation instruction signal sequence according to the number position, and the frequency modulation instruction replacement signal sequence Pt is obtained after replacing the corresponding original signal value 、 .

[0051] The above S104, using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction to obtain a predicted frequency modulation instruction sequence, comprises: Step 1: Replace the signal sequence Pt with the frequency modulation instruction 、 Enter GRU (Gated Recurrent Network) for prediction.

[0052] Replace the signal sequence Pt with the frequency modulation instruction 、 It is divided into two parts according to the length ratio, the first part is used as the input group sequence, and the second part is used as the test group sequence.

[0053] If the length ratio value is 0.9, then the integer sequence [X1, X2, X3....., X 0.9N ] as the input group, followed by an integer sequence of 10% length [X 0.9N+1 ,X 0.9N+2 ,X 0.9N+3 ,...,X N ] as the experimental group.

[0054] At the same time, the unknown group sequence is constructed, with [X N+1 ,X N+2 ,XN+3 ,...,X N+0.1N ] as the unknown group.

[0055] During the prediction process, the input group sequence is used as input to predict the values ​​of the test group and the unknown group.

[0056] Get the first predicted value of the unknown group [X N+1 、 ,X N+2 、 ,X N+3 、 ,...,X N+0.1N 、 ], and the predicted value of the experimental group [X 0.9N+1 、 ,X 0.9N+2 、 ,X 0.9N+3 、 ,...,X N 、 Based on the actual value of the experimental group [X 0.9N+1 ,X 0.9N+2 ,X 0.9N+3 ,...,X N ] and the predicted value [X 0.9N+1 、 ,X 0.9N+2 、 ,X 0.9N+3 、 ,...,X N 、 ] to calculate the first error W1 (mean value of absolute error MAPE) between the two.

[0057] Step 2: Substitute the difference sequence Pt-Pt between the original frequency modulation instruction signal sequence and the frequency modulation instruction replacement signal sequence 、 Enter GRU (Gated Recurrent Network) for prediction.

[0058] The difference sequence Pt-Pt of the original frequency modulation instruction signal sequence and the frequency modulation instruction replacement signal sequence 、 Put it into GRU for prediction, and use the same prediction algorithm as step 1 above to get the second prediction value [X N+1 、、 ,X N+2 、、 ,X N+3 、、 ,...,X N+0.1N 、、 ], the second error W2 of the experimental group.

[0059] Step 3: Fusion the first prediction value, the second prediction value, the first error, and the second error to calculate the final prediction result.

[0060] The fusion algorithm of the final prediction result is:

[0061] Get the predicted frequency modulation instruction sequence [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ].

[0062] When the fusion calculation formula is combined, the ratio of the two is balanced according to the principle of coefficient error.

[0063] The above S105 performs power control of different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence.

[0064] Using the predicted frequency modulation instruction sequence [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ] to conduct power regulation in hybrid energy storage systems. Real-time monitoring of grid frequency deviation, dynamic allocation of power instructions, and rapid suppression of frequency fluctuations.

[0065] With the above Figure 1 Corresponding to the method shown, an embodiment of the present invention also provides a system for predicting energy storage frequency modulation instructions by substitution method, comprising: The command acquisition module acquires the original frequency modulation command signal sequence; The search module finds the strong fluctuation points based on the original frequency modulation command signal sequence; The substitution module replaces the strong fluctuation point signal to generate a substitution signal sequence; The prediction module uses the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction and obtain the predicted frequency modulation instruction sequence; The power control module performs power control of different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence.

[0066] Because the frequency modulation instruction signal sequence has strong volatility and randomness, the traditional prediction method often does not accurately predict the result when it is directly put into the neural network for prediction. In order to solve the shortcomings of strong volatility and randomness of the original sequence, the present invention proposes a new method. First, the fluctuation line is found according to the characteristics of the original sequence, and then it is replaced according to the fluctuation line to form a substitute sequence. The substitute sequence is used to replace the original sequence for prediction, and then the prediction result is corrected to obtain the final prediction result. The method proposed by the present invention can greatly reduce the volatility and randomness of the original sequence and further improve the prediction accuracy.

[0067] In order to further verify the advantages of the present invention, the present invention uses the method of the present invention and the GRU direct prediction method to predict the frequency modulation instruction sequence, and the comparison results are shown in Table 1.

[0068] Table 1: Comparison of performance indicators.

[0069]

[0070] The description of the four evaluation indicators is shown in Table 2.

[0071] Table 2: Description of evaluation indicators.

[0072]

[0073] N represents the sample size, and Represent the actual value and predicted value at time n respectively.

[0074] From the experimental results, we can see that the four evaluation indicators of the algorithm in this application are all reduced, which shows that the proposed method can well improve the prediction accuracy.

[0075] An embodiment of the present application provides a computer device, including: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the computer device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps provided in any embodiment of the present application are performed.

[0076] The computer device provided in the embodiment of the present application includes a processor, a memory, and a bus. The memory is used to store execution instructions, including internal memory and external memory; the internal memory here is also called internal memory, which is used to temporarily store the operation data in the processor and the data exchanged with the external memory such as the hard disk. The processor exchanges data with the external memory through the internal memory. When the electronic device is running, the processor and the memory communicate through the bus, so that the processor executes the following instructions: Obtaining an original frequency modulation command signal sequence; Find the strong fluctuation points based on the original frequency modulation command signal sequence; The strong fluctuation point signal is replaced to generate a replacement signal sequence; Using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction, a predicted frequency modulation instruction sequence is obtained; The power of different energy storage units in the hybrid energy storage system is controlled based on the predicted frequency modulation instruction sequence.

[0077] The embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps provided in any embodiment of the present application are executed. The storage medium can be a volatile or non-volatile computer-readable storage medium.

[0078] Through the description of the above implementation methods, those skilled in the art can clearly understand that the embodiments of the present disclosure can be implemented by hardware, or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the embodiments of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a read-only CD, a USB flash drive, a mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present disclosure.

[0079] Those skilled in the art will appreciate that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes in the accompanying drawings are not necessarily required for implementing the present disclosure.

[0080] Those skilled in the art can understand that the modules in the device in the embodiment can be distributed in the device in the embodiment according to the description of the embodiment, or can be changed accordingly and located in one or more devices different from the present embodiment. The modules in the above embodiment can be combined into one module, or can be further divided into multiple sub-modules.

[0081] The serial numbers of the above-mentioned embodiments of the present disclosure are only for description and do not represent the advantages or disadvantages of the embodiments.

[0082] Obviously, those skilled in the art can make various changes and modifications to the present disclosure without departing from the spirit and scope of the present disclosure. Thus, if these modifications and variations of the present disclosure fall within the scope of the claims of the present disclosure and their equivalents, the present disclosure is also intended to include these modifications and variations.

[0083] Finally, it should be noted that the above description is only an explanation of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail, those skilled in the art can still modify the technical solutions described above or replace some of the technical features with equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting energy storage frequency modulation instructions by substitution method, characterized in that: Includes steps: S101, obtaining an original frequency modulation instruction signal sequence; S102, finding strong fluctuation points based on the original frequency modulation instruction signal sequence; S103, replacing the strong fluctuation point signal to generate a replacement signal sequence; S104, using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction to obtain a predicted frequency modulation instruction sequence; S105. Perform power control of different energy storage units in the hybrid energy storage system based on the predicted frequency modulation instruction sequence.

2. The energy storage frequency modulation instruction prediction method according to claim 1 is characterized in that: In step S101, Dividing the original frequency modulation instruction signal sequence into signal sequences and forming groups; The frequency modulation command signal sequence is Pt=[X1,X2,X3,.X i ..,X N ], the frequency modulation command signal sequence is divided into two parts according to the length ratio, the front part is used as the input group sequence, and the back part is used as the test group sequence.

3. The energy storage frequency modulation instruction prediction method according to claim 2 is characterized in that: If the length ratio is 0.9, then the integer sequence [X1, X2, X3....., X 0.9N ] as the input group, followed by an integer sequence of 10% length [X 0.9N+1 ,X 0.9N+2 ,X 0.9N+3 ,...,X N ] as the test group; construct the unknown group sequence, with [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ] as the unknown group.

4. The energy storage frequency modulation instruction prediction method according to claim 3 is characterized in that: The above S102, finding the strong fluctuation point based on the original frequency modulation instruction signal sequence, includes: Calculating the change slopes and change angle values ​​of adjacent signals, and determining that a number of signal values ​​whose change angle values ​​are greater than a first angle threshold form a first change signal sequence; After performing exponential factorial operation on the first change signal sequence, the change slope and change angle value of the adjacent signals are calculated in the same way, and a number of signal values ​​whose change angle values ​​are greater than the second angle threshold are determined to form a strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ].

5. The energy storage frequency modulation instruction prediction method of the substitution method according to claim 4 is characterized in that: The above S103, replacing the strong fluctuation point signal to generate a replacement signal sequence, includes: Performing a fitting operation on the strongly changing signal sequence to determine a fitting straight line X=Kt+B; Using parameters K and B, we calculate the strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ], each replacement value is inserted into the original frequency modulation instruction signal sequence to obtain the frequency modulation instruction replacement signal sequence Pt 、 .

6. The energy storage frequency modulation instruction prediction method of the substitution method according to claim 5 is characterized in that: The replacement value calculation formula is: 。 7. The energy storage frequency modulation instruction prediction method of the substitution method according to claim 5 is characterized in that: The above S104, using the original frequency modulation instruction signal sequence and the replacement signal sequence to perform prediction to obtain a predicted frequency modulation instruction sequence, comprises: Replace the signal sequence Pt with the frequency modulation instruction 、 Input GRU for prediction and get the first predicted value of the unknown group [X N+1 、 ,X N+2 、 ,X N+3 、 ,...,X N+0.1N 、 ], and the first error W1 of the experimental group was also obtained; The difference sequence Pt-Pt between the original frequency modulation instruction signal sequence and the frequency modulation instruction replacement signal sequence 、 Input GRU for prediction and get the second predicted value of the unknown group [X N+1 、、 ,X N+2 、、 ,X N+3 、、 ,...,X N+0.1N 、、 ], the second error W2 of the experimental group; The first prediction value, the second prediction value, the first error, and the second error are integrated to calculate the final prediction result.

8. The energy storage frequency modulation instruction prediction method of the substitution method according to claim 7 is characterized in that: The fusion algorithm of the final prediction result is: Get the predicted frequency modulation instruction sequence [X N+1 ,X N+2 ,X N+3 ,...,X N+0.1N ].

9. The energy storage frequency modulation instruction prediction method of the substitution method according to claim 5 is characterized in that: For strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] to perform straight line fitting, First, randomly generate a fitting straight line X = Kt + B, and calculate the strong change signal sequence [X 1r ,X 2r ,X 3r ,.X ir ..,X rr ] to this straight line is [d1,d2,d3,..,dr], and the sum of [d1,d2,d3,..,dr] is calculated; The loop is executed to randomly generate M fitting straight lines X=Kt+B, where K and B are both randomly generated; Determine the deviation and the smallest fitting straight line among the M fitting straight lines, and use the fitting straight line y=Kx+B as the selected fitting straight line.

10. The energy storage frequency modulation instruction prediction method of the substitution method according to claim 7 is characterized in that: During the prediction process, the input group sequence is used as input to predict the predicted values ​​of the test group and the unknown group, and the average MAPE of the absolute error is calculated based on the actual value and the predicted value of the test group to generate the first error W1 and the second error W2.

Images