Energy storage system charging and discharging power calculation method, system and device and storage medium

Abstract

The invention discloses an energy storage system charging and discharging power calculation method, system and device and a storage medium. The calculation method comprises the steps of obtaining historical data of an energy storage system and a charging and discharging power daily curve with maximum revenue; constructing sample data according to historical data and the daily curve of the charging and discharging power, and further constructing a sample data set; training, testing and verifying the constructed power calculation model by using the sample data set to obtain a target calculation model; and based on the current meteorological data, the current photovoltaic power generation data, the current energy storage battery data, the current power load data, the meteorological forecast data, the photovoltaic power generation prediction data, the energy storage battery prediction data and the power load prediction data, performing calculation by using the target calculation model to obtain the charging and discharging power of the energy storage system. According to the method, the influence of random fluctuation of the distributed power supply and the user load on the EMS control strategy is reduced, the economic benefit of the industrial and commercial energy storage system is improved, and the project recovery cost period is shortened.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy management, and particularly relates to a method, system, device and storage medium for calculating the charge and discharge power of an energy storage system. Background Art

[0002] Industrial and commercial energy storage is a typical application of the energy storage system on the user side, which can be applied to public facilities such as parks, buildings, factories, etc. Through the coordinated control of the power source, grid, load, and energy storage, it can utilize the peak-valley price difference for arbitrage, suppress the fluctuations of new energy power generation, promote the local consumption of distributed energy, improve the power quality, reduce the demand for transformer capacity, and can also provide power support when the external power grid is out of power to ensure the continuous power supply of important loads.

[0003] In an industrial and commercial energy storage system, the battery management system (BMS), the power conversion system (PCS) of the energy storage, and the energy management system (EMS) coordinate with each other to jointly complete the energy scheduling of the energy storage system. As Figure 1 shown, the battery management system BMS acts as a sensing role, responsible for the monitoring, evaluation, protection, and balancing of the energy storage battery; the power conversion system PCS of the energy storage acts as an execution role, controlling the charging and discharging processes of the energy storage battery and performing the conversion between AC and DC; the energy management system EMS, as the core of the energy storage system, acts as a decision-making role, monitoring and controlling devices such as the photovoltaic inverter, the power conversion system (PCS) of the energy storage, the battery management system (BMS), and the load of the industrial and commercial energy storage system, executing the intelligent EMS control strategy, and improving the reliability and economy of the industrial and commercial energy storage project by means of optimizing the output of distributed power sources, controlling the charge and discharge power of the battery, and switching the load.

[0004] The intelligent energy management and scheduling algorithm is the key to maximizing the benefits of the industrial and commercial energy storage system. At present, the implementation methods of the energy management system EMS include: formulating the charge and discharge plan of the energy storage system in advance by using empirical data combined with time-of-use electricity prices; the method based on multiple linear regression for prediction; the method based on probability statistics, such as Bayesian networks; the method based on deep learning, such as the feedforward artificial neural network, etc.

[0005] A method for formulating a day-ahead charge and discharge plan of an energy storage system according to time-of-use electricity price. The charge and discharge set values of the energy storage system are formulated based on experience, and often cannot quickly respond to the dynamic changes of photovoltaic power and load, resulting in a reduction in the utilization rate of the energy storage system. The multiple linear regression method uses the least squares estimation to jointly predict the optimal combination of multiple related factors. It assumes that there is a linear relationship between the independent variable and the dependent variable. However, the time series of distributed power generation output and electricity load has non-linear characteristics, which cannot be fully expressed by a linear model, resulting in poor prediction accuracy. The Bayesian network based on probability statistics is a conditional probability graph model that describes the conditional dependence relationship between random variables and is suitable for making inferences in imprecise or uncertain knowledge or information, but lacks the consideration of the time correlation of time series data. The method based on the traditional artificial neural network has no memory unit in the network, and the model is unidirectional, only using past information for learning and prediction, without considering the impact of future photovoltaic power generation and electricity load on the energy management control strategy. Summary of the Invention

[0006] The purpose of the present invention is to provide a method, system, device and storage medium for calculating the charge and discharge power of an energy storage system, so as to solve the problem that the traditional energy management system (EMS) is difficult to predict the change trend of photovoltaic power generation and load power within a period of time, and it is difficult to optimize the energy scheduling, resulting in a reduction in the electricity economy of users and an extension of the cost recovery period of energy storage projects.

[0007] The present invention solves the above technical problems through the following technical solutions: A method for calculating the charge and discharge power of an energy storage system includes:

[0008] Obtain the historical data of the energy storage system; wherein, the historical data includes historical meteorological data, historical photovoltaic power generation data, historical energy storage battery data and historical power load data;

[0009] Obtain the daily curve of charge and discharge power that maximizes the revenue of the energy storage system;

[0010] Construct sample data according to the historical data and the daily curve of charge and discharge power, and then construct a sample data set according to the sample data; wherein, the historical data of one day and the corresponding true value of charge and discharge power are used as a sample data.

[0011] Construct a power calculation model;

[0012] Use the sample data set to train, test and verify the power calculation model to obtain a target calculation model;

[0013] Obtain the current meteorological data, current photovoltaic power generation data, current energy storage battery data, current power load data, as well as meteorological forecast data, photovoltaic power generation prediction data, energy storage battery prediction data and power load prediction data;

[0014] Based on the current meteorological data, current photovoltaic power generation data, current energy storage battery data, current power load data, as well as meteorological forecast data, photovoltaic power generation prediction data, energy storage battery prediction data, and power load prediction data, calculate using the target calculation model to obtain the charge and discharge power of the energy storage system.

[0015] Furthermore, the historical meteorological data of each sample data includes the daily maximum temperature, daily minimum temperature, temperature at each collection moment, humidity at each collection moment, and radiation intensity at each collection moment;

[0016] The historical photovoltaic power generation data of each sample data includes the active power of the photovoltaic inverter at each collection moment;

[0017] The historical energy storage battery data of each sample data includes the state of charge of the energy storage battery at each collection moment;

[0018] The historical power load data of each sample data includes the active power of the load at each collection moment.

[0019] Furthermore, use the PyTorch deep learning framework to construct a power calculation model.

[0020] Furthermore, the power calculation model is a bidirectional gated recurrent neural network or a long short-term memory neural network.

[0021] Furthermore, use the sample data set to train the bidirectional gated recurrent neural network, including:

[0022] For each sample data, input the historical data from the 1st to the t-th collection moment into the forward layer of the bidirectional gated recurrent neural network to obtain a forward output; use the historical data from the (t + 1)-th to the n-th collection moment as the backward layer of the bidirectional gated recurrent neural network to obtain a backward output; where n represents the number of collection moments per day;

[0023] Calculate the predicted value of the charge and discharge power at the t-th collection moment according to the forward output and the backward output;

[0024] Calculate the error loss between the predicted value of the charge and discharge power at the t-th collection moment and the true value of the charge and discharge power;

[0025] Adjust the parameters of the bidirectional gated recurrent neural network according to the error loss to achieve network training.

[0026] Furthermore, use the mean square error formula to calculate the error loss between the predicted value of the charge and discharge power at the t-th collection moment and the true value of the charge and discharge power.

[0027] Further, the power load prediction data is determined according to the electricity consumption plan, and the electricity consumption plan includes the production electricity consumption plan and the daily electricity demand.

[0028] Based on the same concept, the present invention further provides an energy storage system, including an energy management system, a battery management system, an energy storage converter, and energy storage batteries; the energy management system is configured to calculate the charging and discharging power according to the charging and discharging power calculation method of the energy storage system as described above, and send the calculated charging and discharging power to the energy storage converter; the energy storage converter is configured to receive the calculated charging and discharging power and control the charging and discharging of the energy storage batteries according to the charging and discharging power.

[0029] Based on the same concept, the present invention further provides an electronic device, including a memory, a processor, and a computer program / instructions stored on the memory, and the processor executes the computer program / instructions to implement the charging and discharging power calculation method of the energy storage system as described above.

[0030] Based on the same concept, the present invention further provides a computer-readable storage medium, on which a computer program / instructions is stored, and when the computer program / instructions is executed by a processor, the charging and discharging power calculation method of the energy storage system as described above is implemented.

[0031] Beneficial effects

[0032] Compared with the prior art, the advantages of the present invention are as follows:

[0033] The present invention uses a bidirectional gated recurrent neural network to learn the non-linear characteristics between the photovoltaic power generation capacity, load information, energy storage regulation ability, and charging and discharging power, and uses the current data and prediction data to perform predictions simultaneously from two directions, and fuses the outputs of the two directions to calculate the charging and discharging power of the energy storage system, effectively capturing the timing relationship in the time series data and the mutual influence between multiple variables, realizing the accurate calculation of the charging and discharging power of the system, and further realizing the optimization of the EMS control strategy; the present invention reduces the influence of the random fluctuations of distributed power sources and user loads on the EMS control strategy, improves the economic benefits of industrial and commercial energy storage systems, and shortens the project cost recovery cycle. Brief description of the drawings

[0034] In order to more clearly illustrate the technical solutions of the present invention, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only one embodiment of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0035] Figure 1 It is the structure diagram of the energy storage system in the background technology of the present invention;

[0036] Figure 2 It is the flowchart of the charge and discharge power calculation method for the energy storage system in the embodiment of the present invention;

[0037] Figure 3 It is the architecture diagram of the bidirectional gated recurrent neural network in the embodiment of the present invention;

[0038] Figure 4 It is the structure diagram of the gated recurrent unit in the embodiment of the present invention. Specific embodiments

[0039] Next, in combination with the accompanying drawings in the embodiments of the present invention, the technical solutions in the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0040] Next, the technical solutions of the present application will be described in detail with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0041] Embodiment 1

[0042] As Figure 2 shown, the charge and discharge power calculation method for the energy storage system provided by the embodiment of the present invention includes the following steps:

[0043] Step 1: Obtain the historical data of the energy storage system.

[0044] The historical data includes historical meteorological data, historical photovoltaic power generation data, historical energy storage battery data, and historical power load data. The historical meteorological data includes daily meteorological data, and the daily meteorological data includes the daily maximum temperature, the daily minimum temperature, the temperature at each collection moment, the humidity at each collection moment, and the radiation intensity at each collection moment. The historical photovoltaic power generation data includes the active power of the photovoltaic inverter at each collection moment; the historical energy storage battery data includes the state of charge of the energy storage battery at each collection moment; the historical power load data includes the active power of the load at each collection moment. In this embodiment, the sampling interval is 1 minute, that is, the temperature, humidity, radiation intensity, the active power of the photovoltaic inverter, the state of charge of the energy storage battery, and the active power of the power load are collected every minute, and the number n of the daily historical data is 24×60 = 1440.

[0045] In order to improve the data quality, the historical data is also preprocessed, and the preprocessing of the present invention includes data cleaning, normalization operations, etc.

[0046] Step 2: Obtain the daily curve of the charge and discharge power that maximizes the benefits of the energy storage system.

[0047] In a specific embodiment of the present invention, a daily curve of charge-discharge power for maximizing the revenue of the energy storage system is calculated in combination with time-of-use electricity prices. To maximize the revenue of the energy storage system, according to different time periods of the time-of-use electricity prices, full-power charging is carried out during the low-valley period, and as fast as possible discharge is carried out during the peak period on the premise of considering grid safety (avoiding reverse power transmission to the grid) and battery balance, so as to calculate the daily curve of charge-discharge power. According to the daily curve of charge-discharge power, the charge-discharge power at each acquisition moment can be obtained.

[0048] Step 3: Construct sample data based on the historical data obtained in Step 1 and the daily curve of charge-discharge power obtained in Step 2, and then construct a sample data set based on all the sample data.

[0049] In this embodiment, the historical data of one day and the corresponding true value of charge-discharge power are used as a sample data. A sample data is a time series data, and a sample data includes data at n acquisition moments. Among them, the data at the t-th acquisition moment includes the daily maximum temperature, the daily minimum temperature, the temperature at the t-th acquisition moment, the humidity at the t-th acquisition moment, the radiation intensity at the t-th acquisition moment, the active power of the photovoltaic inverter at the t-th acquisition moment, the state of charge of the energy storage battery at the t-th acquisition moment, the active power of the load at the t-th acquisition moment, and the true value of charge-discharge power at the t-th acquisition moment. The true value of charge-discharge power at each acquisition moment is determined according to the daily curve of charge-discharge power.

[0050] The sample data set contains multiple sample data, and the sample data set is divided into a training set, a test set, and a validation set by using the cross-validation method.

[0051] Step 4: Construct a power calculation model.

[0052] In a specific embodiment of the present invention, a power calculation model is constructed by using the PyTorch deep learning framework to make the model establishment more flexible. In this embodiment, the power calculation model is a bidirectional gated recurrent neural network (i.e., Bi-GRU model) or a long short-term memory neural network. The basic unit of the long short-term memory neural network consists of an input gate, a forget gate, and an output gate, while the basic unit of the Bi-GRU model consists of an update gate and a reset gate. The structure of the Bi-GRU model is simpler, has fewer parameters, requires fewer training samples, and has a faster training speed.

[0053] Such as Figure 3As shown in the figure, the bidirectional gated recurrent neural network includes an input layer, a forward layer, a backward layer, and an output layer; the input layer is connected to the forward layer and the backward layer, and the forward layer and the backward layer are connected to the output layer. The input layer is used to receive sample data. Both the forward layer and the backward layer include multiple gated recurrent units (GRUs). The forward layer processes the sample data from front to back, and the backward layer processes the sample data from back to front. This bidirectional structure can capture both past and future information, thus more comprehensively modeling the temporal relationship in time series data.

[0054] As Figure 4 shown in the figure, each gated recurrent unit (GRU) is controlled by an update gate and a reset gate to transfer information. The reset gate determines the influence of the data before the t-th acquisition moment on the t-th acquisition moment; the update gate determines how to combine the data at the t-th acquisition moment with the cumulative information before the t-th acquisition moment. Through these gating mechanisms, the Bi-GRU model can adaptively learn the long-term dependencies in time series data and the mutual influence between multiple variables, can well handle the long-term dependence problem of time series data, and at the same time avoid the problem of gradient disappearance. The specific calculation formula of the forward layer is:

[0055] r t = σ(W xr x t + W hr h t-1 + b r )(1)

[0056] z t = σ(W xz x t + W hz h t-1 + b z )(2)

[0057]

[0058] where, x t represents the input vector of the forward layer, h t-1 represents the hidden state of the previous layer, is the candidate hidden state, h t is the output of the forward layer; r t represents the reset gate, z t represents the update gate; σ and tanh represent the sigmoid and tanh activation functions respectively, represents the dot product of matrices; W and b represent the weight matrix and bias corresponding to the gating mechanism and the storage unit respectively.

[0059] Similarly, the output h ′ t of the backward layer can be calculated. The output layer processes the output ht and the output h of the backward layer ′ t are combined to obtain the final output value y t (i.e., the predicted charge-discharge power value at the t-th acquisition moment), and the final output value y t is between [-1, 1]. The specific calculation formula is as follows:

[0060] y t = tanh(W hy [h t , h ′ t +b y )(7)

[0061] Step 5: Use the sample data set to train, test, and validate the power calculation model to obtain the target calculation model.

[0062] Taking the power calculation model as a bidirectional gated recurrent neural network as an example, use each sample data to train the bidirectional gated recurrent neural network, including:

[0063] Step 5.1: Input the historical data from the 1st to the t-th acquisition moment into the forward layer of the bidirectional gated recurrent neural network to obtain the forward output h t ; use the historical data from the (t + 1)-th to the n-th acquisition moment as the backward layer of the bidirectional gated recurrent neural network to obtain the backward output h ′ t ; where n represents the number of acquisition moments per day.

[0064] The historical data at each acquisition moment includes the daily maximum temperature, the daily minimum temperature, the temperature at the corresponding acquisition moment, the humidity at the corresponding acquisition moment, the radiation intensity at the corresponding acquisition moment, the active power of the photovoltaic inverter at the corresponding acquisition moment, the state of charge of the energy storage battery at the corresponding acquisition moment, and the active power of the load at the corresponding acquisition moment.

[0065] Step 5.2: Calculate the predicted charge-discharge power value y t at the t-th acquisition moment according to the forward output h ′ t and the backward output h t .

[0066] Step 5.3: Calculate the error loss between the predicted charge-discharge power value y t at the t-th acquisition moment and the true value of the charge-discharge power.

[0067] In this embodiment, the mean square error formula is used to calculate the error loss between the predicted charge-discharge power value y t at the t-th acquisition moment and the true value of the charge-discharge power, and the mean square error is used to measure the difference between the predicted value and the true value.

[0068] Step 5.4: Adjust the parameters of the bidirectional gated recurrent neural network according to the error loss to achieve network training.

[0069] Update the connection weights in the bidirectional gated recurrent neural network using the backpropagation gradient descent algorithm, and iteratively train repeatedly to gradually learn the features and patterns of the time series data. To enhance robustness, randomly shuffle the sample data in the training set. During the training process, the parameters of the bidirectional gated recurrent neural network are: the number of training iterations epochs is 80, batch_size is 50, the adaptive estimation Adam optimizer is used, and other hyperparameters are set to default values.

[0070] Use the validation set to evaluate the trained bidirectional gated recurrent neural network. The specific evaluation metrics include root mean square error (RMSE) and mean absolute error (MAE). Optimize the bidirectional gated recurrent neural network according to the evaluation results. The performance can be improved by adjusting hyperparameters, changing the model structure, etc. The bidirectional gated recurrent neural network after evaluation and optimization is the target calculation model.

[0071] Step 6: Obtain the current meteorological data, current photovoltaic power generation data, current energy storage battery data, current power load data, as well as meteorological forecast data, photovoltaic power generation prediction data, energy storage battery prediction data, and power load prediction data.

[0072] The current meteorological data is the real-time data collected by the meteorological station. The current photovoltaic power generation data is the active power of the photovoltaic inverter collected in real time. The current energy storage battery data is the state of charge of the energy storage battery collected in real time. The current power load data is the active power of the load collected in real time. The meteorological forecast data comes from the meteorological station. The photovoltaic power generation prediction data is predicted through existing photovoltaic power generation prediction methods. The power load prediction data is determined according to the electricity consumption plan, and the electricity consumption plan includes the production electricity consumption plan and the daily electricity demand. The energy storage battery prediction data is directly filled with specific values. For example, the predicted value of the state of charge of the energy storage battery is 0.5.

[0073] Step 7: Based on the current meteorological data, current photovoltaic power generation data, current energy storage battery data, current power load data, as well as meteorological forecast data, photovoltaic power generation prediction data, energy storage battery prediction data, and power load prediction data, use the target calculation model to calculate and obtain the charge and discharge power of the energy storage system.

[0074] Input the current meteorological data, current photovoltaic power generation data, current energy storage battery data, and current power load data into the forward layer of the target calculation model, and input the meteorological forecast data, photovoltaic power generation prediction data, energy storage battery prediction data, and power load prediction data into the backward layer of the target calculation model. The target calculation model outputs the charge and discharge power of the energy storage system.

[0075] The charge and discharge power calculation method of the present invention based on a bidirectional gated recurrent neural network can make full use of context information, combines the meteorological data at the current moment, the change trends of distributed power sources and electrical loads, and the predicted values for a period of time in the future, and calculates the charge and discharge power of the energy storage system from both the past and future directions. Experimental results show that the charge and discharge power calculation method based on a bidirectional gated recurrent neural network can overcome the adverse factors such as nonlinearity, time-variation, susceptibility to interference, and limited observation time of distributed power sources and electrical load time series, optimizes the performance of the energy management control strategy, improves the economic benefits of industrial and commercial energy storage system projects, and shortens the investment recovery period.

[0076] Embodiment 2

[0077] As Figure 1 shown, the energy storage system provided by the embodiment of the present invention includes an energy management system, a battery management system, an energy storage converter, and an energy storage battery; the energy management system is used to calculate the charge and discharge power according to the charge and discharge power calculation method of the energy storage system in Embodiment 1 of the present invention, and send the calculated charge and discharge power to the energy storage converter; the energy storage converter is used to receive the calculated charge and discharge power and control the charge and discharge of the energy storage battery according to the charge and discharge power, realizing the energy management control of the energy storage system.

[0078] Embodiment 3

[0079] The embodiment of the present invention also provides an electronic device, which includes: a memory, a processor, and a computer program / instructions stored on the memory. The processor executes the computer program / instructions to implement the charge and discharge power calculation method of the energy storage system in Embodiment 1 of the present application.

[0080] Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes according to programs and / or data stored in a read-only memory (ROM) or programs and / or data loaded from a storage section into a random access memory (RAM). The processor can be a multi-core processor or can include multiple processors. In some embodiments, the processor can include a general-purpose main processor and one or more special coprocessors, such as a central processing unit, a graphics processing unit (GPU), a neural network processing unit (NPU), a digital signal processor (DSP), and so on. In the RAM, various programs and data required for device operation are also stored. The processor, ROM, and RAM are connected to each other via a bus. An input / output (I / O) interface is also connected to the bus.

[0081] The above-mentioned processor and memory are jointly used to execute the programs / instructions stored in the memory, and when the programs / instructions are executed by a computer, they can implement the methods, steps, or functions described in the above embodiments.

[0082] Although not shown, an embodiment of the present invention also provides a computer-readable storage medium, on which computer programs / instructions are stored, and when the computer programs / instructions are executed by a processor, the energy storage system charge and discharge power calculation method in the first embodiment of the present application is implemented.

[0083] The readable storage medium includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD), or other optical storage, magnetic cassette tapes, disk storage, or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media do not include transitory computer-readable media, such as modulated data signals and carrier waves.

[0084] The specific embodiments disclosed above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or variations, which should all be covered within the protection scope of the present invention.

Claims

1. A method for calculating the charging and discharging power of an energy storage system, characterized in that: The calculation method includes: Acquire historical data of the energy storage system; wherein the historical data includes historical meteorological data, historical photovoltaic power generation data, historical energy storage battery data and historical power load data; Obtain the daily charging and discharging power curve that maximizes the benefits of the energy storage system; Constructing sample data according to the historical data and the daily charge and discharge power curve, and then constructing a sample data set according to the sample data; wherein one day's historical data and the corresponding real value of the charge and discharge power are taken as a sample data; Construct power calculation model; Using the sample data set to train, test and verify the power calculation model to obtain a target calculation model; Obtain current meteorological data, current photovoltaic power generation data, current energy storage battery data, current power load data, as well as meteorological forecast data, photovoltaic power generation forecast data, energy storage battery forecast data and power load forecast data; Based on current meteorological data, current photovoltaic power generation data, current energy storage battery data, current power load data, as well as meteorological forecast data, photovoltaic power generation prediction data, energy storage battery prediction data and power load prediction data, the target calculation model is used to perform calculations to obtain the charging and discharging power of the energy storage system.

2. The method for calculating the charging and discharging power of the energy storage system according to claim 1, characterized in that: The historical meteorological data of each sample data includes the daily maximum temperature, daily minimum temperature, temperature at each collection time, humidity at each collection time, and radiation intensity at each collection time; The historical photovoltaic power generation data of each sample data includes the active power of the photovoltaic inverter at each collection moment; The historical energy storage battery data of each sample data includes the state of charge of the energy storage battery at each collection moment; The historical power load data of each sample data includes the active power of the load at each collection moment.

3. The method for calculating the charging and discharging power of the energy storage system according to claim 1, characterized in that: Use the PyTorch deep learning framework to build a power calculation model.

4. The method for calculating the charging and discharging power of the energy storage system according to claim 1, characterized in that: The power calculation model is a bidirectional gated recurrent neural network or a long short-term memory neural network.

5. The method for calculating the charging and discharging power of the energy storage system according to claim 4, characterized in that: Using the sample data set to train a bidirectional gated recurrent neural network includes: For each sample data, the historical data from the 1st to the tth collection time is input into the forward layer of the bidirectional gated recurrent neural network to obtain the forward output; the historical data from the t+1th to the nth collection time is used as the backward layer of the bidirectional gated recurrent neural network to obtain the backward output; where n represents the number of collection times per day; Calculating a predicted value of charge and discharge power at the tth collection moment according to the forward output and the backward output; Calculate the error loss between the predicted value of the charge and discharge power at the tth acquisition moment and the actual value of the charge and discharge power; The parameters of the bidirectional gated recurrent neural network are adjusted according to the error loss to achieve network training.

6. The method for calculating the charging and discharging power of the energy storage system according to claim 5, characterized in that: The mean square error formula is used to calculate the error loss between the predicted value of the charge and discharge power at the tth acquisition moment and the actual value of the charge and discharge power.

7. The method for calculating the charging and discharging power of an energy storage system according to any one of claims 1 to 6, characterized in that: The power load forecast data is determined according to a power consumption plan, which includes a production power consumption plan and daily power consumption demand.

8. An energy storage system, comprising an energy management system, a battery management system, an energy storage converter and an energy storage battery; characterized in that: The energy management system is used to calculate the charging and discharging power according to the energy storage system charging and discharging power calculation method described in any one of claims 1 to 7, and send the calculated charging and discharging power to the energy storage inverter; the energy storage inverter is used to receive the calculated charging and discharging power, and control the charging and discharging of the energy storage battery according to the charging and discharging power.

9. An electronic device comprising a memory, a processor, and a computer program / instruction stored in the memory, characterized in that: The processor executes the computer program / instructions to implement the method for calculating the charging and discharging power of the energy storage system according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the method for calculating the charging and discharging power of an energy storage system according to any one of claims 1 to 7 is implemented.

Images