Method and device for determining energy storage capacity of new energy station and computer equipment

Abstract

The invention relates to an energy storage capacity determination method and device of a new energy station and computer equipment, and relates to the technical field of electric power. The method comprises the following steps: based on a target output power prediction model, obtaining predicted output power data of a to-be-constructed first new energy station under predicted meteorological data; determining a fluctuation component which cannot be connected to a grid in the predicted output power data; determining the energy storage capacity of an energy storage system to be configured in the first new energy station based on the fluctuation component which cannot be connected to the grid; wherein the target output power prediction model is obtained by performing fine adjustment on the pre-training output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the pre-training output power prediction model is obtained through pre-training according to a built second new energy station. By adopting the method, the energy storage capacity of the energy storage system to be configured in the new energy station to be constructed can be determined in the planning stage of the new energy station to be constructed.

Description

Technical Field

[0001] The present application relates to the field of electric power technology, and in particular to a method, device, computer equipment, computer-readable storage medium and computer program product for determining energy storage capacity of a new energy station. Background Art

[0002] With the development of new energy, the number of new energy stations, such as wind power stations and photovoltaic stations, is gradually increasing; however, new energy power generation is easily affected by the external environment, thus showing significant volatility and intermittency, which in turn affects the output stability of new energy stations. Therefore, energy storage systems are often configured in new energy stations to ensure the stable output of new energy stations.

[0003] In the related art, the output scenario of the new energy station is usually simulated, and the energy storage capacity of the energy storage system to be configured is determined based on the simulated output scenario.

[0004] However, the simulation of output scenarios often relies on a large amount of historical data of new energy sites, such as historical meteorological data and historical output, which makes it impossible to determine the energy storage capacity of the energy storage system to be configured in the new energy site to be built during the planning stage of the new energy site to be built. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device and computer equipment for determining the energy storage capacity of a new energy station, which can determine the energy storage capacity of the energy storage system to be configured in the new energy station to be constructed during the planning stage of the new energy station to be constructed, in order to address the above-mentioned technical problem that it is impossible to determine the energy storage capacity of the energy storage system to be configured in the new energy station to be constructed during the planning stage of the new energy station to be constructed.

[0006] In a first aspect, the present application provides a method for determining the energy storage capacity of a new energy station, comprising:

[0007] Based on the target output power prediction model corresponding to the first new energy station to be constructed, predicting the output power of the first new energy station under the predicted meteorological data of the first new energy station, to obtain predicted output power data of the first new energy station;

[0008] Determining the fluctuation component of the predicted output power data that cannot be connected to the grid;

[0009] Determining the energy storage capacity of the energy storage system to be configured in the first new energy station based on the statistical characteristics of the fluctuation component that cannot be connected to the grid;

[0010] Among them, the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the first historical output power data is obtained by inverting the first historical meteorological data and the unit operating characteristics of each new energy unit to be configured in the first new energy station; the pre-trained output power prediction model is pre-trained according to the second historical meteorological data and the second historical output power data of the second new energy station that has been constructed.

[0011] In a second aspect, the present application also provides a device for determining energy storage capacity of a new energy station, comprising:

[0012] an output power prediction module, configured to predict the output power of the first new energy station under the predicted meteorological data of the first new energy station based on the target output power prediction model corresponding to the first new energy station to be constructed, and obtain the predicted output power data of the first new energy station;

[0013] A fluctuation component determination module, used to determine the fluctuation component in the predicted output power data that cannot be connected to the grid;

[0014] An energy storage capacity determination module, used to determine the energy storage capacity of the energy storage system to be configured in the first new energy station based on the statistical characteristics of the fluctuation component that cannot be connected to the grid;

[0015] Among them, the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the first historical output power data is obtained by inverting the first historical meteorological data and the unit operating characteristics of each new energy unit to be configured in the first new energy station; the pre-trained output power prediction model is pre-trained according to the second historical meteorological data and the second historical output power data of the second new energy station that has been constructed.

[0016] In a third aspect, the present application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements each step in the above-mentioned method for determining the energy storage capacity of a new energy station when executing the computer program.

[0017] The above-mentioned method, device and computer equipment for determining the energy storage capacity of the new energy station, first, based on the target output power prediction model corresponding to the first new energy station to be constructed, the output power of the first new energy station under the predicted meteorological data of the first new energy station is predicted to obtain the predicted output power data of the first new energy station; then, the fluctuation component that cannot be connected to the grid in the predicted output power data is determined; finally, based on the statistical characteristics of the fluctuation component that cannot be connected to the grid, the energy storage capacity of the energy storage system to be configured in the first new energy station is determined; wherein, the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the first historical output power data is obtained by inverting the first historical meteorological data and the unit operating characteristics of each new energy unit to be configured in the first new energy station; the pre-trained output power prediction model is pre-trained according to the second historical meteorological data and the second historical output power data of the second new energy station that has been constructed. In this way, based on the first historical meteorological data of the first new energy station to be built and the unit operating characteristics of each new energy unit to be configured in the first new energy station, the historical output power of the first new energy station can be inverted to obtain the first historical output power; based on the first historical meteorological data and the first historical output power of the first new energy station, the pre-trained output power prediction model obtained by pre-training according to the second historical meteorological data and the second historical output power of the second new energy station that has been built can be fine-tuned based on the actual situation of the first new energy station, and the target output power prediction model corresponding to the first new energy station can be obtained; based on the predicted meteorological data and the target output power prediction model of the first new energy station, the pre-trained output power prediction model of the second new energy station that has been built can be fine-tuned. A new energy station predicts its output power under predicted meteorological data to obtain predicted output power; based on the fluctuating component of the predicted output power that cannot be connected to the grid, the energy storage capacity of the energy storage system to be configured in the first new energy station can be determined; the method for determining the energy storage capacity of a new energy station based on the above process can invert the historical output power data of the new energy station to be constructed, and then fine-tune the pre-trained output power prediction model to achieve accurate prediction of its output power, thereby determining the energy storage capacity of the energy storage system to be configured. Therefore, the above-mentioned method for determining the energy storage capacity of a new energy station can determine the energy storage capacity of the energy storage system to be configured in the new energy station to be constructed during the planning stage of the new energy station to be constructed. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 A schematic diagram of a flow chart of a method for determining energy storage capacity of a new energy station in one embodiment;

[0020] Figure 2 A flowchart of the steps of fine-tuning a pre-trained output power prediction model to obtain a target output power prediction model in an embodiment;

[0021] Figure 3 A schematic flow chart of the steps of determining the fluctuating component of the predicted output power data that cannot be connected to the grid in one embodiment;

[0022] Figure 4 A flowchart of the steps of determining a sub-signal that cannot be superimposed from the multiple sub-signals by superimposing multiple sub-signals in one embodiment, and determining a fluctuation component that cannot be connected to the grid based on the sub-signal that cannot be superimposed;

[0023] Figure 5 A schematic flow chart of a step of determining whether a superimposed sub-signal satisfies a preset superimposition stop condition in one embodiment;

[0024] Figure 6 A flowchart of steps for performing feature extraction processing on predicted meteorological data to obtain target features of the predicted meteorological data through a plurality of feature extraction networks sequentially connected in a target output power prediction model in one embodiment;

[0025] Figure 7 A schematic flow chart of a method for determining energy storage capacity of a new energy station in another embodiment;

[0026] Figure 8 A schematic flow chart of a method for determining energy storage capacity in an offshore wind farm planning stage in one embodiment;

[0027] Fig. 9 A schematic diagram showing comparison between predicted output power data and actual output power data of an offshore wind farm to be constructed in an embodiment;

[0028] Fig.10 is a schematic diagram of a smoothing component that can be connected to the grid in one embodiment;

[0029] Fig.11 A schematic diagram of a fluctuating component that cannot be connected to the grid in one embodiment;

[0030] Fig.12 It is a structural block diagram of a device for determining energy storage capacity of a new energy station in one embodiment;

[0031] Fig.13 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0033] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0034] In one embodiment, Figure 1 As shown, a method for determining the energy storage capacity of a new energy station is provided. This embodiment uses the method applied to a server as an example for illustration. It can be understood that the method can also be applied to a terminal, and can also be applied to a system including a server and a terminal, and is implemented through the interaction between the server and the terminal; wherein the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services; the terminal can be, but is not limited to, various personal computers, laptops, smart phones, and tablet computers, etc. In this embodiment, the method includes the following steps S102 to S106:

[0035] Step S102, based on the target output power prediction model corresponding to the first new energy station to be constructed, predict the output power of the first new energy station under the predicted meteorological data of the first new energy station to obtain the predicted output power data of the first new energy station.

[0036] The first new energy station is a new energy station to be built. In specific applications, each new energy station has a corresponding new energy type, and the new energy type at least includes a wind power station or a photovoltaic station.

[0037] The predicted meteorological data is the meteorological data of the first new energy station in the future time period, such as wind speed, light intensity, etc. of the first new energy station in the future time period. In a specific application, if the new energy type of the first new energy station is a wind power station, the predicted meteorological data focuses on wind speed and other meteorological data related to wind power generation; if the new energy type of the first new energy station is a photovoltaic station, the predicted meteorological data focuses on light intensity and other meteorological data related to photovoltaic power generation.

[0038] Among them, the target output power prediction model is a neural network model; the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion. Among them, the first historical meteorological data is the meteorological data of the first new energy station in the past time period. Similar to the predicted meteorological data, the focus of the first historical meteorological data is related to the new energy type of the first new energy station. Among them, the first historical output power data is the output power of the first new energy station under the first historical meteorological data; it is easy to understand that the first new energy station is a new energy station to be built, that is, the first new energy station has not yet started construction. Therefore, there is actually no historical output power data for the first new energy station, and then it is necessary to obtain the first historical output power data through inversion.

[0039] Among them, the first historical output power data is obtained by inverting the first historical meteorological data and the unit operation characteristics of each new energy unit to be configured in the first new energy station. Among them, the unit operation characteristics of the new energy unit are used to characterize the corresponding relationship between the new energy and the output power corresponding to the new energy unit; for example, the unit operation characteristics of the wind turbine generator set are used to characterize the corresponding relationship between wind speed and output power, which can be represented by a wind speed-output power curve; for another example, the unit operation characteristics of the photovoltaic generator set are used to characterize the corresponding relationship between light intensity and output power, which can be represented by a light intensity-output power curve.

[0040] Among them, the pre-trained output power prediction model is pre-trained based on the second historical meteorological data and the second historical output power data of the second new energy station that has been built. Among them, the second new energy station is at least one new energy station that has been built, and the new energy type of the second new energy station is the same as the new energy type of the first new energy station. Among them, the second historical meteorological data is the meteorological data of the second new energy station in the past time period. Similar to the predicted meteorological data, the focus of the second historical meteorological data is related to the new energy type of the second new energy station. Among them, the second historical output power data is the output power data under the second historical meteorological data of the second new energy station.

[0041] It is easy to understand that the target output power prediction model corresponding to each first new energy station is related to the first historical meteorological data and the first historical output power data of the new energy station, and the first historical output power data of each new energy station is related to the unit operating characteristics of each new energy unit to be configured in the new energy station, and the first historical meteorological data of each first new energy station and the unit operating characteristics of each new energy unit are different, and then the first historical output power data of each first new energy station is different, and thus the target output power prediction model corresponding to each different first new energy station is different.

[0042] Among them, the predicted output power data is the output power data of the first new energy station under the predicted meteorological data.

[0043] Specifically, for each first new energy station to be constructed, the server first inverts its first historical output power data under the first historical meteorological data based on the first historical meteorological data of the first new energy station and the unit operation characteristics of each new energy unit to be configured therein. Then, based on the first historical meteorological data and the inverted first historical output power data, the server fine-tunes the pre-trained output power prediction model obtained by pre-training the second historical meteorological data and the second historical output power data of at least one second new energy station of the same new energy type as the first new energy station, and obtains the target output power prediction model corresponding to the first new energy unit. Next, the server inputs the predicted meteorological data of the first new energy station into the target output power prediction model, and predicts the output power of the first new energy station under the predicted meteorological data through the target output power prediction model to obtain the predicted output power data.

[0044] Step S104, determining the fluctuation component in the predicted output power data that cannot be connected to the grid.

[0045] Specifically, the server performs signal processing on the predicted output power data, and determines the fluctuation component in the predicted output power data that cannot be connected to the grid according to the grid-connected requirements. This fluctuation component is the part that needs to be smoothed by the energy storage system to be configured. Therefore, the energy storage capacity of the energy storage system to be configured can be determined based on the fluctuation component that cannot be connected to the grid.

[0046] In specific applications, for wind farms, the grid connection requirements require the use of GB / T19963.1-2021 Technical Regulations for Wind Farm Access to Power Systems.

[0047] Step S106, based on the statistical characteristics of the fluctuation component that cannot be connected to the grid, determine the energy storage capacity of the energy storage system to be configured in the first new energy station.

[0048] The statistical features include at least one of a mean, a standard deviation, and a probability distribution.

[0049] Specifically, the server determines the mean and standard deviation of the fluctuating components that cannot be connected to the grid, and determines the probability distribution of the fluctuating components that cannot be connected to the grid based on the mean and standard deviation, and then determines the energy storage compensation amount required for the first new energy station based on the probability distribution, and then determines the energy storage capacity of the energy storage system to be configured in the first new energy station based on the energy storage compensation amount and the continuous response time of the energy storage system to be configured in the first new energy station.

[0050] In the above-mentioned method for determining the energy storage capacity of the new energy station, based on the first historical meteorological data of the first new energy station to be constructed and the unit operating characteristics of each new energy unit to be configured in the first new energy station, the historical output power of the first new energy station can be inverted to obtain the first historical output power; based on the first historical meteorological data and the first historical output power of the first new energy station, the pre-trained output power prediction model obtained by pre-training according to the second historical meteorological data and the second historical output power of the second new energy station that has been constructed can be fine-tuned based on the actual situation of the first new energy station, and the target output power prediction model corresponding to the first new energy station can be obtained; based on the predicted meteorological data and the target output power of the first new energy station The prediction model can predict the output power of the first new energy station under the predicted meteorological data to obtain the predicted output power; based on the fluctuating component of the predicted output power that cannot be connected to the grid, the energy storage capacity of the energy storage system to be configured in the first new energy station can be determined; the method for determining the energy storage capacity of the new energy station based on the above process can invert the historical output power data of the new energy station to be constructed, and then fine-tune the pre-trained output power prediction model to achieve accurate prediction of its output power, thereby determining the energy storage capacity of the energy storage system to be configured. Therefore, the above-mentioned method for determining the energy storage capacity of the new energy station can determine the energy storage capacity of the energy storage system to be configured in the new energy station to be constructed during the planning stage of the new energy station to be constructed.

[0051] In an exemplary embodiment, the unit operation characteristics of each new energy unit are used to characterize the corresponding relationship between the new energy corresponding to the new energy unit and the output power. For example, the unit operation characteristics of a wind generator set are used to characterize the corresponding relationship between wind speed and output power, which can be represented by a wind speed-output power curve; for another example, the unit operation characteristics of a photovoltaic generator set are used to characterize the corresponding relationship between light intensity and output power, which can be represented by a light intensity-output power curve.

[0052] like Figure 2 As shown, the present application also has the following steps for fine-tuning the pre-trained output power prediction model to obtain a target output power prediction model:

[0053] Step S202: based on the unit operation characteristics of each new energy unit and the first historical meteorological data, determine the unit output power data of each new energy unit under the first historical meteorological data.

[0054] Step S204, combining the unit output power data of each new energy unit under the first historical meteorological data to obtain the first historical output power data.

[0055] Step S206: fine-tune the pre-trained output power prediction model based on the first historical meteorological data and the first historical output power data to obtain a target output power prediction model.

[0056] Among them, the unit output power data of each new energy unit includes the output power of the new energy unit at each first historical moment corresponding to the first historical meteorological data; the first historical output power data includes the output power of the first new energy station at each first historical moment corresponding to the first historical meteorological data.

[0057] Specifically, the server first determines the unit operating characteristics of each new energy unit to be configured in the first new energy station, and determines the output power of each new energy unit at each first historical moment corresponding to the first historical meteorological data according to the unit operating characteristics of each new energy unit and the first historical meteorological data, as the unit output power data of each new energy unit under the first historical meteorological data; then, the server superimposes the output power of each new energy unit at each first historical moment to obtain the output power of the new energy station at each first historical moment, as the first historical output power data; finally, the server fine-tunes the pre-trained output power prediction model according to the first historical meteorological data and the first historical output power data, and obtains the target output power prediction model that matches the first new energy station.

[0058] For example, taking the new energy generator set as a wind turbine, for each wind turbine, the server determines the output power of the wind turbine at each first historical moment corresponding to the first historical meteorological data based on the wind speed in the first historical meteorological data and the wind speed-output power curve of the wind turbine.

[0059] Then, for each first historical moment, the server superimposes the output power of each wind turbine generator set at the first historical moment to obtain the output power of the first new energy station at the first historical moment; the output power of the first new energy station at each first historical moment is the first historical output power data.

[0060] In this embodiment, the server can fine-tune the pre-trained output power prediction model to a target output power prediction model that is more consistent with the actual situation of the first new energy station (new energy units to be configured and meteorological conditions) through the unit operating characteristics of each new energy unit to be configured in the first new energy station and the first historical meteorological data of the first new energy station; based on the target output power prediction model, the server can achieve accurate prediction of the output power, thereby determining the energy storage capacity of the energy storage system to be configured.

[0061] In an exemplary embodiment, Figure 3 As shown, the above step S104, determining the fluctuation component in the predicted output power data that cannot be connected to the grid, specifically includes the following steps:

[0062] Step S302: performing modal decomposition processing on the output power signal corresponding to the predicted output power data to obtain a plurality of sub-signals of the output power signal.

[0063] Step S304: determining a sub-signal that cannot be superimposed from the multiple sub-signals by superimposing the multiple sub-signals, and determining a fluctuation component that cannot be connected to the grid based on the sub-signal that cannot be superimposed.

[0064] Among them, the central frequencies of the sub-signals are different.

[0065] Among them, the modal decomposition processing is performed using Ensemble Empirical Mode Decomposition (EEMD).

[0066] Specifically, the server uses EEMD to perform modal decomposition processing on the output power signal corresponding to the predicted output power data to obtain multiple sub-signals with different center frequencies; then, the server superimposes the multiple sub-signals in accordance with the grid-connected requirements and in the order of the corresponding center frequencies from low to high until the superposition is no longer possible; then, after the superposition is no longer possible, the server determines the sub-signals that cannot be superimposed among the multiple sub-signals, and determines the fluctuation components that cannot be connected to the grid based on the sub-signals that cannot be superimposed.

[0067] In this embodiment, the server can decompose the predicted output power data into multiple sub-signals based on EEMD, and based on the superposition processing of the multiple sub-signals, can determine the sub-signals that cannot be superimposed, and further determine the fluctuation components that cannot be connected to the grid.

[0068] In an exemplary embodiment, Figure 4 As shown, the above step 304, by superimposing multiple sub-signals, determines the sub-signals that cannot be superimposed from the multiple sub-signals, and determines the fluctuation component that cannot be connected to the grid based on the sub-signals that cannot be superimposed, specifically includes the following steps:

[0069] Step S402 , arranging the plurality of sub-signals in order of corresponding center frequencies from low to high to obtain a plurality of arranged sub-signals.

[0070] Step S404: determine the first sub-signal among the arranged multiple sub-signals as the first sub-signal to be superimposed, and determine the next sub-signal of the first sub-signal to be superimposed among the arranged multiple sub-signals as the second sub-signal to be superimposed.

[0071] Step S406: superimpose the first sub-signal to be superimposed and the second sub-signal to be superimposed to obtain a superimposed sub-signal.

[0072] Step S408, when the superimposed sub-signal does not meet the preset superposition stop condition, the superimposed sub-signal is determined as a new first sub-signal to be superimposed, and the next sub-signal of the second sub-signal to be superimposed among the arranged multiple sub-signals is determined as a new second sub-signal to be superimposed, and the process of returning to the step of superimposing the first sub-signal to be superimposed and the second sub-signal to be superimposed to obtain the superimposed sub-signal is continued until the obtained superimposed sub-signal meets the preset superposition stop condition.

[0073] Step S410: determine the second to-be-superimposed sub-signal corresponding to the superimposed sub-signal that meets the preset superimposition stop condition, and each sub-signal that is not superimposed among the arranged multiple sub-signals, as sub-signals that cannot be superimposed.

[0074] Step S412, combining the sub-signals that cannot be superimposed to obtain the fluctuation component that cannot be grid-connected.

[0075] Specifically, first, the server arranges multiple sub-signals in order from low to high according to the corresponding center frequencies to obtain multiple arranged sub-signals; then, the server determines the first one of the multiple arranged sub-signals as the first sub-signal to be superimposed, and determines the next one of the multiple arranged sub-signals after the first sub-signal to be superimposed as the second sub-signal to be superimposed.

[0076] Next, the server superimposes the first sub-signal to be superimposed and the second sub-signal to be superimposed to obtain a superimposed sub-signal, and determines whether the superimposed sub-signal satisfies a preset superimposition stop condition.

[0077] If the superimposed sub-signal does not meet the preset superposition stop condition, the server will determine the superimposed sub-signal as a new first sub-signal to be superimposed, determine the next one of the second sub-signal to be superimposed among the arranged multiple sub-signals as a new second sub-signal to be superimposed, and superimpose the new first sub-signal to be superimposed and the new second sub-signal to be superimposed to obtain a new superimposed sub-signal… and repeat this process until the superimposed sub-signal obtains the superposition stop condition.

[0078] If the superimposed sub-signal satisfies the superimposition stopping condition, the server determines the second sub-signal to be superimposed corresponding to the superimposed sub-signal (i.e., the second sub-signal to be superimposed of the superimposed sub-signal) and the remaining non-superimposed sub-signals among the arranged multiple sub-signals as sub-signals that cannot be superimposed.

[0079] Finally, the server sums up the sub-signals that cannot be superimposed to obtain the fluctuation component that cannot be connected to the grid.

[0080] For example, assuming that 16 sub-signals are obtained based on modal decomposition processing; the server arranges multiple sub-signals in order from low to high according to the corresponding center frequency, and superimposes each sub-signal in turn; assuming that when the 9th sub-signal is superimposed, the superimposed sub-signal does not meet the preset superposition stop condition, then the server continues to superimpose the 10th sub-signal; assuming that when the 9th sub-signal is superimposed, the superimposed sub-signal meets the preset superposition stop condition, then the server determines the 9th to 16th sub-signals as sub-signals that cannot be superimposed, and sums the 9th to 16th sub-signals to obtain the fluctuation component that cannot be connected to the grid.

[0081] In this embodiment, the server superimposes each sub-signal according to the center frequency, and determines the sub-signals that cannot be superimposed, and then determines the fluctuating components that cannot be connected to the grid. It is able to separate the fluctuating components that fluctuate violently, do not meet the grid connection requirements, and cannot be connected to the grid from the predicted output power data.

[0082] In an exemplary embodiment, Figure 5 As shown, in the above step S406, after the first to-be-superimposed sub-signal and the second to-be-superimposed sub-signal are superimposed to obtain the superimposed sub-signal, the following steps are specifically further included to determine whether the superimposed sub-signal meets the preset superimposition stop condition:

[0083] Step S502: determining the signal difference between two adjacent sampling points in the superimposed sub-signals.

[0084] Step S504: determining the fluctuation frequency of the superimposed sub-signals according to the number of signal difference values ​​whose corresponding values ​​are greater than the preset fluctuation value in each signal difference value.

[0085] Step S506: When the ratio between the fluctuation frequency and the signal length of the superimposed sub-signal is greater than a preset fluctuation frequency threshold, it is determined that the superimposed sub-signal meets a preset superimposition stop condition.

[0086] Specifically, for the superimposed sub-signal, the server determines the signal difference between two adjacent sampling points in the superimposed sub-signal, and then compares each signal difference with a preset fluctuation value, and counts the number of signal differences whose corresponding values ​​are greater than the preset fluctuation value in each signal difference, and then determines the number as the fluctuation frequency of the superimposed sub-signal.

[0087] Then, the server calculates the ratio between the fluctuation frequency and the signal length of the superimposed sub-signal, and compares the ratio with a preset fluctuation frequency threshold. When the ratio is greater than the preset fluctuation frequency threshold, it is determined that the superimposed sub-signal meets the preset superposition stop condition.

[0088] In this embodiment, the server can obtain the fluctuation frequency of the superimposed sub-signal by comparing the signal difference between two adjacent sampling points in the superimposed sub-signal with a preset fluctuation value, and can determine whether the superimposed sub-signal meets the preset superposition stop condition by comparing the ratio between the fluctuation frequency and the signal length of the signal after superposition with a preset fluctuation frequency threshold.

[0089] In an exemplary embodiment, the above step S106 determines the energy storage capacity of the energy storage system to be configured in the first new energy station based on the statistical characteristics of the fluctuating component that cannot be connected to the grid, and specifically includes the following contents: determining the energy storage compensation amount of the first new energy station based on the statistical characteristics; determining the energy storage capacity of the energy storage system based on the energy storage compensation amount and the continuous response time of the energy storage system.

[0090] The statistical features include at least one of a mean, a standard deviation, and a probability distribution.

[0091] Specifically, the server determines the mean and standard deviation of the fluctuation component that cannot be connected to the grid, and based on the mean and standard deviation, selects an appropriate probability distribution from probability distributions such as normal distribution and gamma distribution, and fits the output power fluctuation data of the first new energy station based on the fluctuation component that cannot be connected to the grid to obtain the energy storage compensation amount of the first new energy station; finally, the server calculates the product of the energy storage compensation amount and the continuous response time of the energy storage system to obtain the energy storage capacity of the energy storage system.

[0092] In specific applications, if the selected probability distribution is normal distribution, then the energy storage compensation ,in, is the standard deviation; if the selected probability distribution is non-normal, then the energy storage compensation is the 97.5th percentile of the fluctuation component; the energy storage capacity of the energy storage system in, is the continuous response time of the energy storage system.

[0093] In this embodiment, the server can fit the output power fluctuation data of the first new energy station through the statistical characteristics of the fluctuation component that cannot be connected to the grid, thereby determining the energy storage capacity of the energy storage system to be configured in the first new energy station.

[0094] In an exemplary embodiment, the above step S102, based on the target output power prediction model corresponding to the first new energy station to be built, predicts the output power of the first new energy station under the predicted meteorological data of the first new energy station to obtain the predicted output power data of the first new energy station, specifically including the following contents: inputting the predicted meteorological data into the target output power prediction model, performing feature extraction processing on the predicted meteorological data through multiple feature extraction networks sequentially connected in the target output power prediction model, and obtaining the target features of the predicted meteorological data; based on the target output power prediction model and the target features, predicting the output power of the first new energy station under the predicted meteorological data to obtain the predicted output power data.

[0095] Among them, there are multiple feature extraction networks connected in sequence in the target output prediction model, and the structures of each feature extraction network are the same.

[0096] Specifically, the server performs feature extraction processing on the predicted meteorological data in sequence through multiple feature extraction networks connected in sequence in the target output power prediction model to obtain target features of the predicted meteorological data; then, based on the target output power prediction model and the target features, the server predicts the output power of the first new energy station under the predicted meteorological data to obtain predicted output power data.

[0097] In this embodiment, the server can extract target features of the predicted meteorological data layer by layer based on multiple feature extraction networks connected in sequence in the target output power prediction model, thereby achieving more accurate output power prediction.

[0098] In an exemplary embodiment, Figure 6 As shown, the above steps, through multiple feature extraction networks sequentially connected in the target output power prediction model, perform feature extraction processing on the predicted meteorological data to obtain the target features of the predicted meteorological data, specifically include the following steps:

[0099] Step S602: In each feature extraction network, a first average feature matrix and a first maximum feature matrix of input information of the feature extraction network are generated, and convolution processing is performed on the first average feature matrix and the first maximum feature matrix respectively to obtain a first convolution-posted average feature matrix and a first convolution-posted maximum feature matrix.

[0100] Step S604, based on the average feature matrix after the first convolution and the maximum feature matrix after the first convolution, determine the channel attention weight matrix of the input information, perform feature extraction processing on the input information based on the channel attention weight matrix, and obtain the first feature of the input information.

[0101] Step S606, generating a second average feature matrix and a second maximum feature matrix of the first feature, and concatenating the second average feature matrix and the second maximum feature matrix to obtain a concatenated feature matrix.

[0102] Step S608, based on the concatenated feature matrix, determine the spatial attention weight matrix of the first feature, extract the first feature based on the spatial attention weight matrix, and obtain the second feature of the input information as the output information of the feature extraction network.

[0103] Step S610: using the output information of the last feature extraction network among the multiple feature extraction networks as the target feature.

[0104] The input information of the first feature extraction network among the multiple feature extraction networks is the predicted meteorological data, and the input information of other feature extraction networks among the multiple feature extraction networks is the output information of the adjacent previous feature extraction network.

[0105] Among them, the feature extraction network is implemented based on the Convolution Block Attention Module (CBAM).

[0106] This embodiment takes the data processing process of a single feature extraction network as an example to illustrate:

[0107] The feature extraction network includes at least a channel attention layer and a spatial attention layer; wherein the channel attention layer includes a global average pooling layer (GAP), a global maximum pooling layer (GMP) and a full connected layer (FC Layer); the spatial attention layer includes a global average pooling layer and a global maximum pooling layer.

[0108] As shown in Formula 1, the server compresses the input information in the spatial dimension based on the global average pooling layer in the channel attention layer to generate the first average feature matrix:

[0109] (Formula 1)

[0110] in, To input information, is the global average pooling operation, is the first average feature matrix.

[0111] As shown in Formula 2, the server extracts the first largest feature matrix of the input information based on the global maximum pooling layer in the channel attention layer:

[0112] (Formula 2)

[0113] in, To input information, is the global maximum pooling operation, is the first largest eigenvalue matrix.

[0114] The server performs feature transformation processing on the first average feature matrix and the first maximum feature matrix based on the fully connected layer. Specifically, a convolutional neural network (CNN) is used as the fully connected layer. As shown in Formula 3, the server performs feature transformation processing on the first average feature matrix and the first maximum feature matrix based on the fully connected layer to obtain the first convolution average feature matrix and the first convolution maximum feature matrix:

[0115] (Formula 3)

[0116] in, For full connection operation, is the convolution operation, is the weight matrix of the convolution operation, is the weight matrix of the fully connected operation, is the linear rectification activation function.

[0117] As shown in Formula 4, the server determines the channel attention weight matrix of the input information based on the average feature matrix after the first convolution and the maximum feature matrix after the first convolution:

[0118] (Formula 4)

[0119] in, is the channel attention weight matrix, is the sigmoid activation function.

[0120] As shown in Formula 5, the server performs feature extraction processing on the input information based on the channel attention weight matrix to obtain the first feature of the input information:

[0121] (Formula 5)

[0122] As shown in Formula 6, the server compresses the first feature in the spatial dimension based on the global average pooling layer in the spatial attention layer to generate a second average feature matrix:

[0123] (Formula 6)

[0124] in, is the second average feature matrix.

[0125] As shown in Formula 7, the server extracts the second largest feature matrix of the first feature based on the global maximum pooling layer in the spatial attention layer:

[0126] (Formula 7)

[0127] in, is the second largest eigenvalue matrix.

[0128] Then, the server concatenates the second average feature matrix and the second maximum feature matrix to obtain a concatenated feature matrix .

[0129] As shown in Formula 8, the server determines the spatial attention weight matrix of the first feature based on the concatenated feature matrix:

[0130] (Formula 8)

[0131] in, is the spatial attention weight matrix.

[0132] As shown in Formula 9, the server performs feature extraction processing on the first feature based on the spatial attention weight matrix to obtain the second feature of the input information:

[0133] (Formula 9)

[0134] in, The second feature.

[0135] In this embodiment, the server can extract the target features of the predicted meteorological data layer by layer through the feature extraction networks connected in sequence, and the channel attention mechanism and spatial attention mechanism in each feature extraction network, thereby achieving more accurate output power prediction.

[0136] In an exemplary embodiment, each feature extraction network also includes a convolution layer, a batch normalization layer and a ReLU activation function; the input information of each feature extraction network needs to first pass through the convolution layer, the batch normalization layer and the ReLU activation function in sequence to generate preliminary features, and then enter the channel attention layer and the spatial attention layer to obtain output information.

[0137] In an exemplary embodiment, both the pre-training process of the pre-trained output power prediction model and the fine-tuning process of the target output power prediction model adopt an early stopping mechanism to prevent overfitting caused by over-training and save the time cost of calculation.

[0138] In an exemplary embodiment, the present application also pre-processes the predicted meteorological data, the first historical meteorological data, the second historical meteorological data, and the second historical output power data; the pre-processing process is as follows:

[0139] For the data to be preprocessed (the predicted meteorological data, the first historical meteorological data, the second historical meteorological data, and the second historical output power data), the mean and standard deviation of the data to be preprocessed are determined, and based on the mean and standard deviation of the data to be preprocessed, the standard score (Z-score value) of each data in the data to be preprocessed is determined; then, the server determines the data whose corresponding standard score is greater than the outlier threshold as outlier data; then, the server determines the null value data and outlier data in the data to be preprocessed as data to be interpolated; then, for each data to be interpolated, the server determines the target data corresponding to the data to be interpolated based on the two known data before and after the data to be interpolated, and replaces the data to be interpolated with the target data.

[0140] As shown in Formula 10, the server calculates the target data using the linear interpolation formula:

[0141] (Formula 10)

[0142] in, is the sequence number of the data to be interpolated in the data to be preprocessed, is the number of data to be interpolated, For the The target data corresponding to the data to be interpolated, For the The previous data of the data to be interpolated, For the The next data to be interpolated, For the The timestamp of the data to be interpolated, For the The timestamp of the previous data to be interpolated.

[0143] In an exemplary embodiment, the forecast meteorological data, the first historical meteorological data, the second historical meteorological data and the second historical output power data in the present application are all stored in a CSV (Comma-Separated Values) file, and the reading process for the CSV file is as follows: read the first 10,000 bytes of the data file, automatically detect the encoding format of the file, and output the encoding type and confidence of the detection result; try to read the CSV file according to the encoding format detected in the previous step. If it fails, try other common encoding formats in turn until it is successfully read or an error is thrown; after the data is read successfully, perform a Max-Min normalization operation to map the read data to the range of [0, 1].

[0144] In an exemplary embodiment, if Figure 7 As shown, another method for determining the energy storage capacity of a new energy station is provided, and the method is applied to a server as an example for explanation, including the following steps:

[0145] Step S702: pre-training a pre-trained output power prediction model based on the second historical meteorological data and the second historical output power data of the second new energy station that has been constructed.

[0146] Step S704, based on the unit operation characteristics of each new energy unit to be configured in the first new energy station to be constructed and the first historical meteorological data, determine the unit output power data of each new energy unit under the first historical meteorological data.

[0147] Step S706, combining the unit output power data of each new energy unit under the first historical meteorological data to obtain the first historical output power data of the first new energy station.

[0148] Step S708, based on the first historical meteorological data and the first historical output power data of the first new energy station, fine-tune the pre-trained output power prediction model to obtain a target output power prediction model corresponding to the first new energy station.

[0149] Step S710: Based on the target output power prediction model, the output power of the first new energy station under the predicted meteorological data of the first new energy station is predicted to obtain the predicted output power data of the first new energy station.

[0150] Step S712, performing modal decomposition processing on the output power signal corresponding to the predicted output power data to obtain multiple sub-signals of the output power signal, and superimposing the multiple sub-signals in sequence until the obtained superimposed sub-signal meets the preset superposition stop condition, and determining the fluctuation component that cannot be connected to the grid based on the sub-signals that cannot be superimposed.

[0151] Step S714, based on the statistical characteristics of the fluctuation component that cannot be connected to the grid, determine the energy storage compensation amount of the first new energy station, and based on the energy storage compensation amount and the continuous response time of the energy storage system, determine the energy storage capacity of the energy storage system.

[0152] In this embodiment, the method for determining the energy storage capacity of a new energy station based on the above process can invert the historical output power data of the new energy station to be built, and then fine-tune the pre-trained output power prediction model to achieve accurate prediction of its output power, thereby determining the energy storage capacity of the energy storage system to be configured. Therefore, the above method for determining the energy storage capacity of a new energy station can determine the energy storage capacity of the energy storage system to be configured in the new energy station to be built during the planning stage of the new energy station to be built.

[0153] In order to more clearly illustrate the method for determining the energy storage capacity of a new energy station provided in the embodiment of the present application, the method for determining the energy storage capacity of the new energy station is specifically described below with a specific embodiment, but it should be understood that the embodiment of the present application is not limited thereto. Figure 8 As shown, in one exemplary embodiment, the present application also provides a method for determining energy storage capacity in the offshore wind farm planning stage, which specifically includes the following steps:

[0154] 1. Obtain the historical meteorological data and historical output power data of the offshore wind farms that have been put into operation, as well as the historical meteorological data of the offshore wind farms to be built, and perform preprocessing.

[0155] 2. Based on the historical meteorological data and historical output power data of the offshore wind farms that have been put into operation, pre-train the offshore wind power prediction model. Among them, the pre-trained offshore wind power prediction model integrates the convolutional neural network CNN and the convolutional attention module CBAM.

[0156] 3. Inverse the historical output power data of the offshore wind farm to be constructed based on the historical meteorological data of the offshore wind farm to be constructed and the operating characteristics of the wind turbines to be deployed.

[0157] 4. Based on the historical meteorological data and historical output power data of the offshore wind farm to be constructed, the pre-trained offshore wind power prediction model is fine-tuned to obtain the target offshore wind power prediction model of the offshore wind farm to be constructed.

[0158] 5. According to the target offshore wind power prediction model and predicted meteorological data of the offshore wind farm to be constructed, predict the predicted output power data of the offshore wind farm to be constructed. Fig. 9 Shown is a schematic diagram comparing the predicted output power data and the actual output power data of the offshore wind farm to be constructed.

[0159] 6. Use the empirical set modal decomposition to deeply decompose the predicted output power data to obtain a series of power sub-signals with different central frequencies. Consider the requirements of wind power grid connection to determine the fluctuation components that cannot be connected to the grid. For the fluctuation components that cannot be connected to the grid, it is necessary to use the energy storage system to smooth them. Therefore, the energy storage capacity of the energy storage system to be configured is determined according to the probability statistical characteristics of the fluctuation components that cannot be connected to the grid. Fig.10 The schematic diagram of the smoothing component that can be connected to the grid is shown as Fig.11 Shown is a schematic diagram of the fluctuation component that cannot be connected to the grid.

[0160] In this embodiment, firstly, by integrating the convolutional neural network CNN and the convolutional attention module CBAM, the model's perception and learning ability of time series features is enhanced; CBAM accurately captures important features through the synergy of the channel attention layer and the spatial attention layer, and improves the model's feature expression ability and prediction accuracy. Secondly, through the "pre-training-fine-tuning" strategy: the model is pre-trained using the data of the offshore wind farm that has been put into operation, and the general wind power output characteristics are learned. Then, the data of the offshore wind farm to be built is used for fine-tuning, and the model parameters are adjusted to adapt to the characteristics of the offshore wind farm to be built. This strategy not only accelerates the convergence speed of the model, but also significantly improves the adaptability and prediction performance of the model, and reduces the risk of overfitting. Thirdly, the predicted meteorological data is deeply decomposed through the empirical set mode decomposition, and the intrinsic modes of different frequencies are extracted. By gradually superimposing IMFs and evaluating the signal change ratio, the high-frequency fluctuation components that cannot be connected to the grid are scientifically determined, and the smooth components that can participate in the grid connection and the high-volatility components that cannot participate in the grid connection are effectively separated, providing a solid foundation for the accurate calculation of energy storage capacity. Finally, combined with the statistical characteristics of the fluctuation component, the required capacity of the energy storage system is scientifically calculated. The above method not only improves the scientificity and rationality of energy storage configuration, but also optimizes the utilization efficiency of energy storage resources and reduces the system operation cost.

[0161] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0162] Based on the same inventive concept, the embodiment of the present application also provides a device for determining the energy storage capacity of a new energy station for implementing the above-mentioned method for determining the energy storage capacity of a new energy station. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in the embodiments of the device for determining the energy storage capacity of one or more new energy stations provided below can be referred to the limitations of the method for determining the energy storage capacity of a new energy station above, and will not be repeated here.

[0163] In an exemplary embodiment, Fig.12 As shown, a device for determining energy storage capacity of a new energy station is provided, comprising: an output power prediction module 1202, a fluctuation component determination module 1204 and an energy storage capacity determination module 1206, wherein:

[0164] The output power prediction module 1202 is used to predict the output power of the first new energy station under the predicted meteorological data of the first new energy station based on the target output power prediction model corresponding to the first new energy station to be built, and obtain the predicted output power data of the first new energy station.

[0165] The fluctuation component determination module 1204 is used to determine the fluctuation component that cannot be connected to the grid in the predicted output power data.

[0166] The energy storage capacity determination module 1206 is used to determine the energy storage capacity of the energy storage system to be configured in the first new energy station based on the statistical characteristics of the fluctuation component that cannot be connected to the grid.

[0167] Among them, the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the first historical output power data is obtained by inverting the first historical meteorological data and the unit operation characteristics of each new energy unit to be configured in the first new energy station; the pre-trained output power prediction model is pre-trained according to the second historical meteorological data and the second historical output power data of the second new energy station that has been built.

[0168] In an exemplary implementation, the unit operation characteristics of each new energy unit are used to characterize the corresponding relationship between the new energy corresponding to the new energy unit and the output power.

[0169] The energy storage capacity determination device of the new energy station also includes a model prediction model fine-tuning module, which is used to determine the unit output power data of each new energy unit under the first historical meteorological data based on the unit operation characteristics of each new energy unit and the first historical meteorological data; combine the unit output power data of each new energy unit under the first historical meteorological data to obtain the first historical output power data; based on the first historical meteorological data and the first historical output power data, fine-tune the pre-trained output power prediction model to obtain the target output power prediction model.

[0170] In an exemplary implementation, the fluctuation component determination module 1204 is also used to perform modal decomposition processing on the output power signal corresponding to the predicted output power data to obtain multiple sub-signals of the output power signal; the center frequency of each sub-signal is different; through superposition processing of multiple sub-signals, sub-signals that cannot be superimposed are determined from the multiple sub-signals, and based on the sub-signals that cannot be superimposed, the fluctuation component that cannot be connected to the grid is determined.

[0171] In an exemplary implementation, the fluctuation component determination module 1204 is further used to arrange the multiple sub-signals in the order of corresponding center frequencies from low to high to obtain the multiple arranged sub-signals; determine the first sub-signal in the multiple arranged sub-signals as the first sub-signal to be superimposed, and determine the next sub-signal of the first sub-signal to be superimposed in the multiple arranged sub-signals as the second sub-signal to be superimposed; superimpose the first sub-signal to be superimposed and the second sub-signal to be superimposed to obtain the superimposed sub-signal; if the superimposed sub-signal does not meet the preset superposition stop condition, determine the superimposed sub-signal as a new first sub-signal to be superimposed, determine the next sub-signal of the second sub-signal to be superimposed in the multiple arranged sub-signals as a new second sub-signal to be superimposed, and return to the step of superimposing the first sub-signal to be superimposed and the second sub-signal to be superimposed to obtain the superimposed sub-signal until the obtained superimposed sub-signal meets the preset superposition stop condition; determine the second sub-signal to be superimposed corresponding to the superimposed sub-signal that meets the preset superposition stop condition, and each sub-signal that is not superimposed in the multiple arranged sub-signals as sub-signals that cannot be superimposed; combine the sub-signals that cannot be superimposed to obtain the fluctuation component that cannot be connected to the grid.

[0172] In an exemplary implementation, the fluctuation component determination module 1204 is also used to determine the signal difference between two adjacent sampling points in the superimposed sub-signal; determine the fluctuation frequency of the superimposed sub-signal based on the number of signal difference values ​​whose corresponding values ​​are greater than the preset fluctuation value in each signal difference; and determine that the superimposed sub-signal satisfies the preset superposition stop condition when the ratio between the fluctuation frequency and the signal length of the superimposed sub-signal is greater than the preset fluctuation frequency threshold.

[0173] In an exemplary implementation, the energy storage capacity determination module 1206 is further used to determine the energy storage compensation amount of the first new energy station based on statistical characteristics; and determine the energy storage capacity of the energy storage system based on the energy storage compensation amount and the continuous response time of the energy storage system.

[0174] In an exemplary implementation, the output power prediction module 1202 is also used to input the predicted meteorological data into the target output power prediction model, perform feature extraction processing on the predicted meteorological data based on multiple feature extraction networks sequentially connected in the target output power prediction model, and obtain the target features of the predicted meteorological data; based on the target output power prediction model and the target features, predict the output power of the first new energy station under the predicted meteorological data to obtain predicted output power data.

[0175] In an exemplary implementation, the output power prediction module 1202 is also used to generate a first average feature matrix and a first maximum feature matrix of the input information of the feature extraction network in each feature extraction network, and perform convolution processing on the first average feature matrix and the first maximum feature matrix respectively to obtain a first convolution average feature matrix and a first convolution maximum feature matrix; the input information of the first feature extraction network in the multiple feature extraction networks is the predicted meteorological data, and the input information of other feature extraction networks in the multiple feature extraction networks is the output information of the adjacent previous feature extraction network; based on the first convolution average feature matrix and the first convolution maximum feature matrix, determine Determine the channel attention weight matrix of the input information, perform feature extraction on the input information based on the channel attention weight matrix, and obtain the first feature of the input information; generate the second average feature matrix and the second maximum feature matrix of the first feature, concatenate the second average feature matrix and the second maximum feature matrix to obtain the concatenated feature matrix; determine the spatial attention weight matrix of the first feature based on the concatenated feature matrix, perform feature extraction on the first feature based on the spatial attention weight matrix, and obtain the second feature of the input information as the output information of the feature extraction network; use the output information of the last feature extraction network among the multiple feature extraction networks as the target feature.

[0176] Each module in the energy storage capacity determination device of the above-mentioned new energy station can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0177] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig.13As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data of the new energy station. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for determining the energy storage capacity of a new energy station is implemented.

[0178] Those skilled in the art will understand that Fig.13 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0179] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above-mentioned method embodiments when executing the computer program.

[0180] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0181] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the above method embodiments are implemented.

[0182] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0183] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0184] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for determining the energy storage capacity of a new energy station, characterized in that: The method comprises: Based on the target output power prediction model corresponding to the first new energy station to be constructed, predicting the output power of the first new energy station under the predicted meteorological data of the first new energy station, to obtain predicted output power data of the first new energy station; Determining the fluctuation component of the predicted output power data that cannot be connected to the grid; Determining the energy storage capacity of the energy storage system to be configured in the first new energy station based on the statistical characteristics of the fluctuation component that cannot be connected to the grid; Among them, the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the first historical output power data is obtained by inverting the first historical meteorological data and the unit operating characteristics of each new energy unit to be configured in the first new energy station; the pre-trained output power prediction model is pre-trained according to the second historical meteorological data and the second historical output power data of the second new energy station that has been constructed.

2. The method according to claim 1, characterized in that The unit operation characteristics of each new energy unit are used to characterize the corresponding relationship between the new energy corresponding to the new energy unit and the output power; The target output power prediction model is obtained by: Based on the unit operation characteristics of each new energy unit and the first historical meteorological data, determining the unit output power data of each new energy unit under the first historical meteorological data; Combining the unit output power data of each new energy unit under the first historical meteorological data to obtain the first historical output power data; Based on the first historical meteorological data and the first historical output power data, the pre-trained output power prediction model is fine-tuned to obtain the target output power prediction model.

3. The method according to claim 1, characterized in that: The determining of the fluctuation component of the predicted output power data that cannot be connected to the grid includes: Performing modal decomposition processing on the output power signal corresponding to the predicted output power data to obtain multiple sub-signals of the output power signal; each sub-signal has a different central frequency; By superimposing the multiple sub-signals, a sub-signal that cannot be superimposed is determined from the multiple sub-signals, and based on the sub-signal that cannot be superimposed, the fluctuation component that cannot be connected to the grid is determined.

4. The method according to claim 3, characterized in that: The method of determining a sub-signal that cannot be superimposed from the multiple sub-signals by superimposing the multiple sub-signals, and determining the fluctuation component that cannot be connected to the grid based on the sub-signal that cannot be superimposed, includes: Arranging the plurality of sub-signals in order of corresponding center frequencies from low to high to obtain a plurality of arranged sub-signals; Determine the first sub-signal among the arranged multiple sub-signals as a first sub-signal to be superimposed, and determine the next sub-signal of the first sub-signal to be superimposed among the arranged multiple sub-signals as a second sub-signal to be superimposed; Superimposing the first to-be-superimposed sub-signal and the second to-be-superimposed sub-signal to obtain a superimposed sub-signal; In the case where the superimposed sub-signal does not meet the preset superposition stop condition, the superimposed sub-signal is determined as the new first sub-signal to be superimposed, the next sub-signal of the second sub-signal to be superimposed among the arranged multiple sub-signals is determined as the new second sub-signal to be superimposed, and the step of returning to superimpose the first sub-signal to be superimposed and the second sub-signal to be superimposed to obtain the superimposed sub-signal is performed until the obtained superimposed sub-signal meets the preset superposition stop condition; Determine the second to-be-superimposed sub-signal corresponding to the superimposed sub-signal that meets the preset superimposition stop condition, and each sub-signal that is not superimposed in the plurality of arranged sub-signals, as the sub-signals that cannot be superimposed; The sub-signals that cannot be superimposed are combined to obtain the fluctuation component that cannot be grid-connected.

5. The method according to claim 4, characterized in that After superimposing the first to-be-superimposed sub-signal and the second to-be-superimposed sub-signal to obtain a superimposed sub-signal, the method further includes: Determine the signal difference between two adjacent sampling points in the superimposed sub-signals; Determining the fluctuation frequency of the superimposed sub-signals according to the number of signal difference values ​​whose corresponding values ​​are greater than the preset fluctuation value in each signal difference; When the ratio between the fluctuation frequency and the signal length of the superimposed sub-signal is greater than a preset fluctuation frequency threshold, it is determined that the superimposed sub-signal satisfies the preset superposition stop condition.

6. The method according to any one of claims 1 to 5, characterized in that: The determining, based on the statistical characteristics of the fluctuation component that cannot be connected to the grid, the energy storage capacity of the energy storage system to be configured in the first new energy station comprises: Based on the statistical characteristics, determining the energy storage compensation amount of the first new energy station; The energy storage capacity of the energy storage system is determined based on the energy storage compensation amount and the continuous response time of the energy storage system.

7. The method according to any one of claims 1 to 5, characterized in that: The target output power prediction model corresponding to the first new energy station to be constructed is used to predict the output power of the first new energy station under the predicted meteorological data of the first new energy station to obtain the predicted output power data of the first new energy station, including: Inputting the predicted meteorological data into the target output power prediction model, performing feature extraction processing on the predicted meteorological data through a plurality of feature extraction networks sequentially connected in the target output power prediction model, and obtaining target features of the predicted meteorological data; Based on the target output power prediction model and the target characteristics, the output power of the first new energy station under the predicted meteorological data is predicted to obtain the predicted output power data.

8. The method according to claim 7, characterized in that The method of performing feature extraction processing on the predicted meteorological data through a plurality of feature extraction networks sequentially connected in the target output power prediction model to obtain target features of the predicted meteorological data includes: In each feature extraction network, a first average feature matrix and a first maximum feature matrix of input information of the feature extraction network are generated, and convolution processing is performed on the first average feature matrix and the first maximum feature matrix respectively to obtain a first convolution average feature matrix and a first convolution maximum feature matrix; the input information of the first feature extraction network among the multiple feature extraction networks is the predicted meteorological data, and the input information of other feature extraction networks among the multiple feature extraction networks is the output information of the adjacent previous feature extraction network; Determine a channel attention weight matrix of the input information based on the first average feature matrix after convolution and the first maximum feature matrix after convolution, and perform feature extraction processing on the input information based on the channel attention weight matrix to obtain a first feature of the input information; Generate a second average feature matrix and a second maximum feature matrix of the first feature, and concatenate the second average feature matrix and the second maximum feature matrix to obtain a concatenated feature matrix; Based on the concatenated feature matrix, determining a spatial attention weight matrix of the first feature, performing feature extraction on the first feature based on the spatial attention weight matrix, and obtaining a second feature of the input information as output information of the feature extraction network; The output information of the last feature extraction network among the multiple feature extraction networks is used as the target feature.

9. A device for determining energy storage capacity of a new energy station, characterized in that: The device comprises: an output power prediction module, configured to predict the output power of the first new energy station under the predicted meteorological data of the first new energy station based on the target output power prediction model corresponding to the first new energy station to be constructed, and obtain the predicted output power data of the first new energy station; A fluctuation component determination module, used to determine the fluctuation component in the predicted output power data that cannot be connected to the grid; An energy storage capacity determination module, used to determine the energy storage capacity of the energy storage system to be configured in the first new energy station based on the statistical characteristics of the fluctuation component that cannot be connected to the grid; Among them, the target output power prediction model is obtained by fine-tuning the pre-trained output power prediction model according to the first historical meteorological data of the first new energy station and the first historical output power data obtained by inversion; the first historical output power data is obtained by inverting the first historical meteorological data and the unit operating characteristics of each new energy unit to be configured in the first new energy station; the pre-trained output power prediction model is pre-trained according to the second historical meteorological data and the second historical output power data of the second new energy station that has been constructed.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

Images