Super-capacity energy storage grid connection method and system

Abstract

The invention discloses a super-capacitance energy storage grid connection method and system, and belongs to the technical field of super-capacitance energy storage, and the method comprises the steps: firstly obtaining a current frequency modulation instruction sequence, and calculating a fluctuation coefficient of the current frequency modulation instruction sequence; and then, inputting the fluctuation coefficient into a pre-trained error adjustment model to predict and obtain a final internal resistance increment. And if the internal resistance increment exceeds a set first preset value, the system automatically calls the lithium battery for compensation. In the training process of the error adjustment model, a fluctuation coefficient of a historical frequency modulation instruction sequence and an internal resistance increment sequence are used as a training set, and an internal resistance fluctuation function is constructed and optimized until the average error rate is lower than a second preset value, so that the accuracy of the model is ensured. The method aims at improving the performance and stability of the super-capacitance energy storage system in power grid frequency modulation.

Description

Technical Field

[0001] This invention belongs to the field of supercapacity energy storage technology, and relates to a supercapacity energy storage grid connection method and system. Background Technology

[0002] In power systems, thermal power units serve as the primary source of electricity generation, and their frequency regulation capabilities are crucial for maintaining the stability and reliability of the power grid. Traditional frequency regulation methods often rely on the inherent regulation capabilities of thermal power units themselves. However, due to the relatively slow response speed of thermal power units, they struggle to quickly respond to rapid changes in grid frequency. To compensate for this deficiency, hybrid energy storage systems (typically composed of supercapacitors and lithium batteries) are widely used to assist thermal power units in frequency regulation.

[0003] In hybrid energy storage systems, supercapacitors and lithium batteries each perform different frequency regulation tasks. Specifically, supercapacitors, due to their high power density and rapid charge / discharge capabilities, are primarily responsible for frequency regulation commands in the high-frequency range, while lithium batteries, with their high energy density, are better suited for frequency regulation in the low-frequency range. This division of labor effectively improves the frequency regulation performance of thermal power units.

[0004] However, in practical applications, the traditional method of using hybrid energy storage to assist the frequency regulation of thermal power units has significant shortcomings. This method simply transmits the difference between the frequency regulation command and the thermal power unit to the hybrid energy storage system, with the supercapacitor handling the high-frequency portion and the lithium battery handling the low-frequency portion, but it ignores the impact of high-frequency fluctuations on the internal resistance of the supercapacitor.

[0005] The increased internal resistance of supercapacitors due to high-frequency fluctuations can trigger a series of negative effects. First, with the increase in internal resistance, the energy loss during the charging and discharging process of the supercapacitor will also increase accordingly. This means that, under the same conditions, the energy that a supercapacitor can store will decrease, i.e., its energy density will decrease. This not only reduces the overall efficiency of the energy storage system but may also affect the stable operation of the power grid.

[0006] Secondly, increased internal resistance also negatively impacts the power performance of supercapacitors. During discharge, capacitors with higher internal resistance require more energy to overcome the voltage drop caused by the internal resistance, resulting in a reduction in actual output power. This poses a significant challenge for high-frequency modulation tasks that demand rapid response.

[0007] Furthermore, a long-term increase in internal resistance can accelerate the aging process of the supercapacitor's internal materials, leading to a shorter lifespan. This not only increases maintenance costs but may also threaten the safe and stable operation of the power grid. Simultaneously, increased internal resistance may also increase the risk of damage to the capacitor under extreme conditions such as overcharging and over-discharging, further exacerbating the instability of its application.

[0008] It is worth noting that existing control technologies do not fully consider the specific relationship between the internal resistance of the supercapacitor and frequency fluctuations. This makes it difficult to accurately predict and assess the impact of high-frequency fluctuations on the performance of supercapacitors in practical applications, thus hindering the development of effective control measures to optimize the frequency regulation performance of hybrid energy storage systems. Summary of the Invention

[0009] The purpose of this invention is to solve the technical problem in the prior art that the failure to consider the specific relationship between the internal resistance of the supercapacitor and frequency fluctuations leads to a decrease in the performance of the supercapacitor and large fluctuations in the frequency modulation command, and to provide a supercapacitor energy storage grid connection method and system.

[0010] To achieve the above objectives, the present invention employs the following technical solution: The first aspect of this invention provides a method for grid-connecting supercapacitor energy storage, comprising the following steps: Obtain the current frequency modulation command sequence; Calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence; The fluctuation coefficient is input into the pre-trained error adjustment model to obtain the final internal resistance increment; If the final internal resistance increment is greater than the first preset value, then the lithium battery compensation will be activated. The pre-trained error adjustment model is trained using the following method: The fluctuation coefficient sequence of historical frequency modulation command sequences and the internal resistance increment sequence corresponding to historical frequency modulation commands are used as training sets. Construct an internal resistance fluctuation function that characterizes the relationship between internal resistance fluctuation and fluctuation coefficient; initialize the internal resistance fluctuation function; The error adjustment model is trained using a training set. During training, the fluctuation coefficient sequence is substituted into the internal resistance fluctuation function, and the average error rate is calculated by combining it with the internal resistance increment sequence. If the average error rate is greater than the second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model.

[0011] Furthermore, the calculation of the fluctuation coefficient of the current frequency modulation command sequence specifically involves: Construct an approximation function for the current frequency modulation command sequence; initialize the approximation function; Based on the current frequency modulation command sequence and the approximation function, calculate the approximation error sequence. When the mean of the approximation error sequence is greater than the preset mean, update the approximation function. Iterate this step until the mean of the approximation error is less than the preset mean, and obtain the final approximation function and the number of iterations. Find the first derivative of the final approximation function to obtain the first approximation derivative; calculate the number of solutions when the first approximation derivative is zero; The fluctuation coefficient is obtained based on the number of solutions and the number of iterations when the first-order approximation derivative is zero.

[0012] Furthermore, the approximation function after initialization for:

[0013] Among them, the highest digit coefficient Second highest coefficient and last coefficient All are random values; This represents the value at each position in the current frequency modulation command sequence.

[0014] Furthermore, the approximation error sequence is as follows:

[0015] in, For the first The approximation error sequence of each iteration; The current frequency modulation command sequence is the i The value of the position; For the first During the next iteration The approximate function value; ; The length of the current frequency modulation command sequence.

[0016] Furthermore, the mean of the approximation error sequence is described as follows:

[0017] in, For the first The mean of the approximation error sequence in each iteration.

[0018] Furthermore, the updated approximation function is specifically as follows:

[0019] in, For the first The approximation function for the next iteration; ; ; ; In the formula, For random number functions; Indicates the first The approximation function of the nth iteration +1 digit coefficient; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; express of Power of 1 It is the Euler number; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position.

[0020] Furthermore, the internal resistance fluctuation function is described as follows:

[0021] in, For the first The internal resistance fluctuation function of the next cycle Represents the number of loops; ; ; ; ; ; ; In the formula, Indicates the average error. , For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; M The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence; For activation functions; `max()` is a random function; `max()` is a maximum value function. For the first The supplementary coefficient for the next cycle; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle; For the first The average error rate of the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle.

[0022] Furthermore, the average error rate is specifically as follows:

[0023] in, For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence.

[0024] A second aspect of the present invention provides a computer storage medium storing instructions that, when executed, enable the supercapacity energy storage grid connection method to be implemented.

[0025] A third aspect of the present invention provides an ultracapacity energy storage grid-connected system, comprising: The data acquisition module is used to acquire the current frequency modulation command sequence; The fluctuation coefficient calculation module is used to calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence. The internal resistance measurement module is used to input the fluctuation coefficient into the pre-trained error adjustment model to obtain the final internal resistance increment; If the final internal resistance increment of the grid-connected module is greater than the first preset value, the lithium battery will be activated for compensation. The pre-trained error adjustment model is trained using the following method: The fluctuation coefficient sequence of historical frequency modulation command sequences and the internal resistance increment sequence corresponding to historical frequency modulation commands are used as training sets. Construct an internal resistance fluctuation function that characterizes the relationship between internal resistance fluctuation and fluctuation coefficient; initialize the internal resistance fluctuation function; The error adjustment model is trained using a training set. During training, the fluctuation coefficient sequence is substituted into the internal resistance fluctuation function, and the average error rate is calculated by combining it with the internal resistance increment sequence. If the average error rate is greater than the second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model.

[0026] Compared with the prior art, the present invention has the following beneficial effects: This invention discloses a method for grid-connected supercapacitor energy storage. Through a pre-trained error adjustment model, it can accurately predict and adjust the fluctuation values ​​generated by frequency regulation commands, ensuring that the final fluctuation value remains within a preset range, thereby effectively improving the stability of the supercapacitor energy storage grid-connected system. When the final fluctuation value exceeds the preset value, this invention can intelligently utilize lithium batteries for compensation, avoiding energy waste and equipment damage caused by excessive fluctuations, and optimizing energy utilization efficiency. During the training process of the error adjustment model, by calculating the fluctuation coefficient, detecting the internal resistance increment, and iteratively calculating until the average error rate meets the preset conditions, the model can more accurately reflect the actual grid state, improving the model's prediction accuracy and robustness. By introducing the calculation of approximation error and fluctuation coefficient, this invention can more comprehensively consider the impact of frequency regulation commands on the grid, enhancing the system's adaptability to different frequency regulation commands and improving the system's flexibility and reliability.

[0027] Furthermore, the proposed supercapacitive energy storage grid-connected system utilizes a pre-trained error adjustment model in its fluctuation calculation module to accurately predict grid fluctuations caused by frequency regulation commands and calculate the final fluctuation value. This mechanism effectively reduces the risk of grid instability due to acquisition errors, improving grid connection accuracy and system stability. When the final fluctuation value exceeds a preset threshold, the grid-connected module can quickly mobilize lithium batteries for compensation, ensuring that grid fluctuations are controlled within a safe range. This intelligent compensation strategy not only avoids the negative impact of grid fluctuations on system stability but also optimizes energy allocation and improves energy utilization efficiency. Attached Figure Description

[0028] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a block diagram of the supercapacity energy storage grid connection method of the present invention; Figure 2 This is a block diagram of the supercapacity energy storage grid-connected system of the present invention.

[0030] Among them, 201-data acquisition module; 202-fluctuation coefficient calculation module; 203-internal resistance measurement module; 204-grid connection module. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and marked in the accompanying drawings can generally be arranged and designed in various different configurations.

[0032] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0033] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0034] The present invention will now be described in further detail with reference to the accompanying drawings: See Figure 1 One embodiment of the present invention discloses a method for grid connection of supercapacity energy storage, comprising the following steps: S1, Obtain the current frequency regulation command sequence; obtain the current frequency regulation command sequence from the power grid dispatch center. These commands typically include information such as power grid frequency deviation, target adjustment amount, and time requirements.

[0035] S2, Calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence; S3, input the fluctuation coefficient into the pre-trained error adjustment model to obtain the final internal resistance increment; The pre-trained error adjustment model is trained using the following method: The fluctuation coefficient sequence of historical frequency modulation command sequences and the internal resistance increment sequence corresponding to historical frequency modulation commands are used as training sets. Construct an internal resistance oscillation function with internal resistance oscillation and oscillation coefficient; initialize the internal resistance oscillation function; The fluctuation coefficient sequence is substituted into the internal resistance fluctuation function and combined with the internal resistance increment sequence to calculate the average error rate. If the average error rate is greater than the second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model.

[0036] S4. If the final internal resistance increment is greater than the first preset value, then the lithium battery compensation is activated.

[0037] This embodiment acquires the frequency regulation command sequence from the power grid dispatch center in real time and calculates its fluctuation coefficient, enabling precise capture of the dynamic changes in power grid frequency. Combined with a pre-trained error adjustment model, this method accurately predicts the internal resistance increment caused by acquisition errors, thereby achieving fine-grained control over the charging and discharging process of the supercapacitor energy storage system. This not only improves the response speed of power grid frequency regulation but also significantly enhances its accuracy, contributing to the stability of the power grid frequency. In conjunction with a lithium battery compensation mechanism, this invention automatically activates the lithium battery for compensation when the final internal resistance increment exceeds a preset value, ensuring that the energy storage system always remains in optimal operating condition. This not only achieves intelligent scheduling of the energy storage system but also reduces unnecessary energy loss, thus achieving energy conservation and consumption reduction.

[0038] One embodiment of the present invention provides a method for grid connection of supercapacity energy storage, specifically: S1, obtain the current frequency modulation command sequence; S2, Calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence; S201, Construct an approximation function for the current frequency modulation command sequence; initialize the approximation function; the initialized approximation function is:

[0039] Among them, the highest digit coefficient Second highest coefficient and last coefficient All values ​​are random.

[0040] S202, based on the current frequency modulation command sequence and the approximation function, calculate the approximation error sequence. When the mean of the approximation error sequence is greater than the preset mean, update the approximation function. Iterate this step until the mean of the approximation error is less than the preset mean, and obtain the final approximation function and the number of iterations. The approximation error sequence is as follows:

[0041] in, For the first The approximation error sequence of the next iteration; The current frequency modulation command sequence is the i The value of the position; For the first During the next iteration The approximate function value; ; The length of the current frequency modulation command sequence.

[0042] The mean of the approximation error sequence is described as follows:

[0043] in, For the first The mean of the approximation error sequence in each iteration. The updated approximation function is specifically:

[0044] in, For the first The approximation function of the next iteration; the highest-order coefficient It is a random value; ; ; ; In the formula, For random number functions; Indicates the first The approximation function of the nth iteration +1 digit coefficient; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; express of Power of 1 It is the Euler number; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position.

[0045] S203, take the first derivative of the final approximation function to obtain the first approximation derivative, and calculate the number of intersections between the first approximation derivative and the X-axis (that is, the number of solutions when the first approximation derivative is zero). S204, based on the number of intersections and the number of iterations, yields the fluctuation coefficient.

[0046] S3, input the fluctuation coefficient into the pre-trained error adjustment model to obtain the final internal resistance increment; S301, the fluctuation coefficient sequence of the historical frequency modulation command sequence and the internal resistance increment sequence corresponding to the historical frequency modulation command are used as the training set; S302, Construct the internal resistance ripple function with internal resistance ripple and ripple coefficient; Initialize the internal resistance ripple function; The internal resistance fluctuation function is described as follows:

[0047] in, Internal resistance ripple function Represents the number of loops; ; ; ; ; ; ; In the formula, Indicates the average error. , For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; M The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence; For activation functions; `max()` is a random function; `max()` is a maximum value function. For the first The supplementary coefficient for the next cycle; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle; For the first The average error rate of the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle.

[0048] S303, the error adjustment model is trained using the training set. During training, the fluctuation coefficient sequence is substituted into the internal resistance fluctuation function and combined with the internal resistance increment sequence to calculate the average error rate. If the average error rate is greater than a second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model. The calculation of the average error rate specifically involves:

[0049] in, For the first The corresponding iteration of the 1st cycle The internal resistance increment during the second frequency modulation period, here is the first The frequency modulation period refers to each frequency modulation corresponding to the historical frequency modulation command sequence of the training set. The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence.

[0050] S4. If the final internal resistance increment is greater than the first preset value, then the lithium battery compensation is activated.

[0051] One embodiment of the present invention provides a method for grid connection of supercapacity energy storage, specifically: S1, let pci=[ , , ,..., [This refers to the instructions handled by the supercapacitor during a certain frequency modulation process; initialization iteration count] =1.

[0052] Let the initial approximation function be: ; , and All The coefficients at time, where, These are the highest-order coefficients of the approximation function in the first iteration; These are the second-highest coefficients of the approximation function in the first iteration; These are the last coefficients of the approximation function in the first iteration; , and All are given randomly. Substitute get ,Bundle Substitute get ,Bundle Substitute get , ,Bundle Substitute get ; The length of the current frequency modulation command sequence; This is the value of the first bit in the current frequency modulation command sequence; This is the value of the second bit in the current frequency modulation command sequence; This is the value of the 3rd bit in the current frequency modulation command sequence; The current frequency modulation command sequence is the The value of the bit; This is the approximation function value of the first bit of the current frequency modulation command sequence during the first iteration; This is the approximation function value of the second bit of the current frequency modulation command sequence during the first iteration; This is the approximation function value of the third bit of the current frequency modulation command sequence during the first iteration; For the first iteration, the current frequency modulation command sequence is... The approximation function value of the bit value.

[0053] The approximation error sequence of the first iteration for:

[0054] Based on this, the mean of the approximation error sequence for the first iteration is calculated. : .

[0055] If the mean of the approximation error sequence in the first iteration is If the value is greater than the preset mean of 0.2, then the approximation function is updated. The approximation function in the second iteration is then:

[0056] in, For random values, ; ; ; These are the highest-order coefficients of the approximation function during the second iteration; These are the second-highest coefficients of the approximation function during the second iteration; These are the third high-order coefficients of the approximation function during the second iteration; These are the last coefficients of the approximation function during the second iteration; This represents the approximation function during the second iteration; It is a random function; It is the Euler number.

[0057] Next, put Substitute get ,Bundle Substitute get ,Bundle Substitute get ,......,Bundle Substitute get ; This is the approximation function value of the first bit of the current frequency modulation command sequence during the second iteration; This is the approximation function value of the second bit of the current frequency modulation command sequence during the second iteration; This is the approximation function value of the third bit of the current frequency modulation command sequence during the second iteration; For the second iteration, the current frequency modulation command sequence is... The approximation function value of the bit value.

[0058] Then, the approximation error sequence of the second iteration for:

[0059] Find the mean of this set of approximation error sequences:

[0060] If the mean of the approximation error sequence in the second iteration is If the value is greater than the preset mean of 0.2, then the approximation function is updated. In the third iteration, the approximation function is:

[0061] in, For random values, ;

[0062] ; ; In the formula, This represents the approximation function in the 3rd iteration; These are the highest-order coefficients of the approximation function during the third iteration; These are the second-highest coefficients of the approximation function during the third iteration; This represents the third highest-order coefficient of the approximation function during the third iteration. This represents the fourth highest coefficient of the approximation function during the third iteration; These are the last coefficients of the approximation function during the third iteration.

[0063] Next, put Substitute get ,Bundle Substitute get ,Bundle Substitute get , ,Bundle Substitute get ;in, This is the approximation function value of the first bit of the current frequency modulation command sequence during the third iteration; This is the approximation function value of the second bit of the current frequency modulation command sequence during the third iteration; This is the approximation function value of the third bit of the current frequency modulation command sequence during the third iteration; For the third iteration, the current frequency modulation command sequence is... The approximation function value of the bit value.

[0064] The approximation error sequence of the third iteration for:

[0065] Based on this, the mean of the approximation error sequence for the third iteration is: .

[0066] Following the above method, repeat the process in a loop until the [number]th ... The approximation function of the next iteration The corresponding number The mean of the approximation error sequence in the next iteration If the value is less than 0.2, the number of iterations is obtained. The number of iterations As one of the influencing factors of volatility coefficient.

[0067] For the final number The approximation function of the next iteration Find the derivative to obtain the first derivative. Let the number of intersections between this first derivative and the x-axis be (i.e., the number of points when the first derivative is 0). The number of solutions is M Then the volatility coefficient .

[0068] S2: Collect 50 consecutive overcapacitance frequency modulation commands pc1, pc2, ..., pci, ..., pc50, and calculate the corresponding 50 fluctuation coefficients according to the method in S1. , , ,..., And detect 50 increments of the internal resistance value during the corresponding single frequency modulation period. , , ... Then, based on the fluctuation coefficient and the internal resistance increment, an internal resistance fluctuation function is constructed, specifically: First, let the initial form of the describing function (i.e., the internal resistance ripple function of the first cycle) be:

[0069] in, , , and The initial values ​​are all , Indicates the first The compensation value for the next loop is initially set to 0; , , ,..., Substitute respectively The average error of the first cycle was further calculated. The average error rate of the first cycle ; For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; For the first The fluctuation coefficient corresponding to the second frequency modulation, that is, the first The fluctuation coefficient corresponding to the overcapacitance frequency modulation command; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle.

[0070] like If the value is greater than 5%, then the internal resistance fluctuation function is corrected to obtain the internal resistance fluctuation function for the second cycle:

[0071] in, ; ; ; ; ; ; In the formula, Indicates the first The compensation value of the next loop; , , ,..., Substitute respectively To further obtain the first The average error of the next cycle and the Average error rate of the next cycle ; For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; For the first The fluctuation coefficient corresponding to the second frequency modulation, that is, the first The fluctuation coefficient corresponding to the overcapacitance frequency modulation command; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle.

[0072] like If the value is greater than 5%, the internal resistance fluctuation function is corrected to obtain the internal resistance fluctuation function for the third cycle:

[0073] in, ; ;

[0074]

[0075]

[0076] ; In the formula, It is an activation function; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle; This represents the internal resistance fluctuation function of the third cycle.

[0077] Repeat the above process until a certain average error rate is reached. Less than 5%, forming the final internal resistance fluctuation function:

[0078] in, Internal resistance ripple function Represents the number of loops; ; ; ; ; ; ; In the formula, Indicates the average error. ; For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; M The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence; For activation functions; `max()` is a random function; `max()` is a maximum value function. For the first The supplementary coefficient for the next cycle; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle; For the first The average error rate of the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle.

[0079] After S3 forms the final internal resistance fluctuation function, each time a current frequency modulation command arrives, the fluctuation coefficient of the current frequency modulation command is first calculated, and then based on... Calculate the final internal resistance increment, which is equal to the internal resistance fluctuation function value divided by the internal resistance value at the time of the last frequency modulation command. If the final internal resistance increment is greater than the first preset value of 2%, then a portion of the command is compensated by the battery.

[0080] It should be noted that, in actual implementation, all parameters are processed by extremum normalization / Min-MaxScaling before calculation.

[0081] One embodiment of the present invention provides a computer storage medium storing instructions that, when executed on a computer or similar device, enable the implementation of the supercapacity energy storage grid-connected method as described above. The computer storage medium in this embodiment can be any physical device or virtual space capable of storing data for computer reading. For example, it can be a hard disk drive, solid-state drive, memory module, read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, flash memory card, optical disk, or any other form of computer-readable storage medium. In this embodiment, the instructions for the supercapacity energy storage grid-connected method are stored in the aforementioned computer storage medium. These instructions can exist in the form of machine code, assembly code, source code or object code written in a high-level programming language (such as C, C++, Java, Python, etc.). When a computer or similar device reads and executes these instructions, they trigger a series of operations to implement the supercapacity energy storage grid-connected method. These operations include, but are not limited to: The system reads the frequency modulation command and inputs it into the pre-trained error adjustment model; calculates and outputs the final fluctuation value; determines whether the final fluctuation value is greater than the preset fluctuation value, and decides whether to adjust the lithium battery for compensation accordingly; if compensation is required, it calculates and outputs the compensation command to the lithium battery control system; and implements all steps in the error adjustment model training process, including calculating the fluctuation coefficient, detecting the internal resistance increment, and iteratively calculating the fluctuation value.

[0082] See Figure 2 One embodiment of the present invention provides an extracapacity energy storage grid-connected system, comprising: Data acquisition module 201 is used to acquire the current frequency modulation command sequence; The fluctuation coefficient calculation module 202 is used to calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence; The internal resistance measurement module 203 is used to input the fluctuation coefficient into the pre-trained error adjustment model to obtain the final internal resistance increment; If the final internal resistance increment of the grid-connected module 204 is greater than the first preset value, the lithium battery compensation will be activated. The pre-trained error adjustment model is trained using the following method: The fluctuation coefficient sequence of historical frequency modulation command sequences and the internal resistance increment sequence corresponding to historical frequency modulation commands are used as training sets. Construct an internal resistance oscillation function with internal resistance oscillation and oscillation coefficient; initialize the internal resistance oscillation function; The fluctuation coefficient sequence is substituted into the internal resistance fluctuation function and combined with the internal resistance increment sequence to calculate the average error rate. If the average error rate is greater than the second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model.

[0083] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for grid connection of supercapacity energy storage, characterized in that, Includes the following steps: Obtain the current frequency modulation command sequence; Calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence; The fluctuation coefficient is input into the pre-trained error adjustment model to obtain the final internal resistance increment; If the final internal resistance increment is greater than the first preset value, then the lithium battery compensation will be activated. The pre-trained error adjustment model is trained using the following method: The fluctuation coefficient sequence of historical frequency modulation command sequences and the internal resistance increment sequence corresponding to historical frequency modulation commands are used as training sets. Construct an internal resistance fluctuation function that characterizes the relationship between internal resistance fluctuation and fluctuation coefficient; initialize the internal resistance fluctuation function; The error adjustment model is trained using a training set. During training, the fluctuation coefficient sequence is substituted into the internal resistance fluctuation function, and the average error rate is calculated by combining the internal resistance increment sequence. If the average error rate is greater than the second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model.

2. The method for connecting supercapacity energy storage to the grid according to claim 1, characterized in that, The calculation of the fluctuation coefficient of the current frequency modulation command sequence is specifically as follows: Construct an approximation function for the current frequency modulation command sequence; initialize the approximation function; Based on the current frequency modulation command sequence and the approximation function, calculate the approximation error sequence. When the mean of the approximation error sequence is greater than the preset mean, update the approximation function. Iterate this step until the mean of the approximation error is less than the preset mean, to obtain the final approximation function and the number of iterations; Find the first derivative of the final approximation function to obtain the first approximation derivative; calculate the number of solutions when the first approximation derivative is zero; The fluctuation coefficient is obtained based on the number of solutions and the number of iterations when the first-order approximation derivative is zero.

3. The method for connecting supercapacity energy storage to the grid according to claim 2, characterized in that, The approximation function after initialization for: Among them, the highest digit coefficient Second highest coefficient and last coefficient All are random values; This represents the value at each position in the current frequency modulation command sequence.

4. The method for connecting supercapacity energy storage to the grid according to claim 2, characterized in that, The approximation error sequence is as follows: in, For the first The approximation error sequence of the next iteration; The current frequency modulation command sequence is the i The value of the position; For the first During the next iteration The approximate function value; ; The length of the current frequency modulation command sequence.

5. The method for connecting supercapacity energy storage to the grid according to claim 4, characterized in that, The mean of the approximation error sequence is described as follows: in, For the first The mean of the approximation error sequence in each iteration.

6. The method for connecting supercapacity energy storage to the grid according to claim 5, characterized in that, The updated approximation function is specifically: in, For the first The approximation function for the next iteration; ； ； ； In the formula For random number functions; Indicates the first The approximation function of the nth iteration +1 digit coefficient; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; express of Power of 1 It is the Euler number; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position; Indicates the first The approximation function of the nth iteration The coefficient of the position.

7. The method for connecting supercapacity energy storage to the grid according to claim 1, characterized in that, The internal resistance fluctuation function is described as follows: in, For the first The internal resistance fluctuation function of the next cycle Represents the number of loops; ； ； ； ； ； ； In the formula, Indicates the average error. , For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; M The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence; For activation functions; `max()` is a random function; `max()` is a maximum value function. For the first The supplementary coefficient for the next cycle; For the first The supplementary coefficient for the next cycle; For the first Correction coefficient for the next cycle; For the first The compensation coefficient for the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; For the first The compensation value for the next cycle; For the first The average error rate of the next cycle; For the first The correction term for the next iteration; For the first The compensation term for the next cycle; middle This represents the volatility coefficient.

8. The method for connecting supercapacity energy storage to the grid according to claim 7, characterized in that, The average error rate is specifically: in, For the first The corresponding iteration of the 1st cycle The internal resistance increment during the secondary frequency modulation period; The length of the fluctuation coefficient sequence of the historical frequency modulation command sequence.

9. A computer storage medium storing instructions, characterized in that, When the instruction is executed, the supercapacity energy storage grid connection method according to any one of claims 1-8 is implemented.

10. A supercapacity energy storage grid-connected system, based on the supercapacity energy storage grid-connected method according to any one of claims 1 to 8, characterized in that, include: The data acquisition module is used to acquire the current frequency modulation command sequence; The fluctuation coefficient calculation module is used to calculate the fluctuation coefficient of the current frequency modulation command sequence based on the current frequency modulation command sequence. The internal resistance measurement module is used to input the fluctuation coefficient into the pre-trained error adjustment model to obtain the final internal resistance increment; If the final internal resistance increment of the grid-connected module is greater than the first preset value, the lithium battery will be activated for compensation. The pre-trained error adjustment model is trained using the following method: The fluctuation coefficient sequence of historical frequency modulation command sequences and the internal resistance increment sequence corresponding to historical frequency modulation commands are used as training sets. Construct an internal resistance fluctuation function that characterizes the relationship between internal resistance fluctuation and fluctuation coefficient; initialize the internal resistance fluctuation function; The error adjustment model is trained using a training set. During training, the fluctuation coefficient sequence is substituted into the internal resistance fluctuation function, and the average error rate is calculated by combining the internal resistance increment sequence. If the average error rate is greater than the second preset value, the internal resistance fluctuation function is corrected until the average error rate is less than the second preset value, thus obtaining the trained error adjustment model.

Images