An overload imbalance automatic adjustment compensation charging and discharging energy storage system

Abstract

The present application relates to the field of electric energy storage, and specifically to an overload and imbalance automatic adjustment compensation charging and discharging energy storage system. It uses artificial intelligence technology based on deep learning to monitor and analyze the state of the power grid, electricity price information and energy storage battery status, and mines out the time series change information of the power grid status, electricity price information and battery status, and then intelligently generates charging and discharging strategies based on the dependency relationship between the three. In this way, dynamic adjustment of the energy storage system can be achieved, effectively compensating for the overload and imbalance problems of the power grid, improving equipment utilization, and ensuring the stable operation of the power system and the quality of power service for users.

Description

Technical Field

[0001] The present application relates to the field of electric energy storage, and in particular to an overload and imbalance automatic adjustment and compensation charging and discharging energy storage system. Background Art

[0002] With the continuous growth of global energy demand and the increasing openness of the power market, the complexity and operation requirements of the power system are constantly increasing. Traditional power systems face many challenges in dealing with peak loads, balancing supply and demand, and optimizing resource allocation. In particular, with the popularization of distributed generation technologies, such as large-scale access to wind power and photovoltaic power generation, the volatility of the power grid has increased significantly, bringing tremendous pressure to the stability and security of the power system.

[0003] The current industrial and commercial energy storage systems take advantage of the difference in peak and valley electricity prices of the power grid, and arbitrage to obtain economic benefits by charging or double charging when the electricity price is low, and discharging or implementing a double discharge strategy during peak hours. However, the charging and discharging control methods of these energy storage systems are relatively simple, and they can only be connected to the power grid for operation when the capacity of the energy storage system reaches a certain scale (such as energy storage substations). For ordinary industrial and commercial users, their energy storage systems can only be consumed internally and cannot be connected to the grid. This will lead to a decrease in the utilization rate of the energy storage system when the user load is significantly reduced.

[0004] Therefore, in order to improve the utilization rate and economic benefits of the energy storage system, an optimized overload imbalance automatic adjustment compensation charging and discharging energy storage system is expected. Summary of the invention

[0005] This application is made in view of the above problems. One object of this application is to provide an overload imbalance automatic adjustment compensation charging and discharging energy storage system.

[0006] The embodiment of the present application provides an overload imbalance automatic adjustment compensation charging and discharging energy storage system, which includes:

[0007] An input / output port, the input / output port being used to be electrically connected to a power grid so as to absorb electric energy from the power grid or release electric energy to the power grid through the input / output port;

[0008] A power conversion system (PCS) for converting the AC power of the grid into DC power for storage or converting the stored DC power back into the AC power and supplying it to the grid or load;

[0009] Energy storage batteries, used to store and release electrical energy;

[0010] A battery management system, used to monitor and manage the charging and discharging process of the energy storage battery;

[0011] The controller is used to determine the charging and discharging strategy based on the grid status, electricity price information and battery status.

[0012] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the controller includes:

[0013] A power grid information acquisition module, used to acquire the time series of the electricity price information, the time series of the power grid state and the time series of the battery state, wherein the power grid state includes a voltage value, a current value and a frequency value, and the battery state includes a remaining power, a battery voltage value, a battery current value and a SOC;

[0014] An electricity price information time series encoding module, used for performing time series encoding on the time series of the electricity price information to obtain an electricity price time series associated implicit feature vector;

[0015] A multi-parameter time series correlation coding module, used for performing time series correlation coding of state parameters on the time series of the power grid state and the time series of the battery state to obtain a time series correlation characteristic diagram of the power grid state and a time series correlation characteristic diagram of the battery state;

[0016] A feature autocorrelation reinforcement module, used to perform autocorrelation attention reinforcement on the grid state time series correlation feature graph and the battery state time series correlation feature graph respectively to obtain a reinforced grid state time series correlation feature vector and a reinforced battery state time series correlation feature vector;

[0017] A charge and discharge strategy generation module is used to determine the charge and discharge strategy based on the probabilistic dependency relationship between the enhanced grid state time series associated feature vector, the enhanced battery state time series associated feature vector and the electricity price time series associated implicit feature vector, wherein the charge and discharge strategy is used to indicate starting charging or starting discharging.

[0018] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the electricity price information timing encoding module is used to:

[0019] The time series of the electricity price information is input into a sequence encoder based on RNN to obtain the electricity price time series associated implicit feature vector.

[0020] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the multi-parameter timing association encoding module includes:

[0021] A power grid state multi-parameter time series association encoding unit, used for arranging the time series of the power grid state into a power grid state parameter time series joint matrix according to the parameter sample dimension and the time dimension, and inputting the power grid state parameter time series joint matrix into a power grid state feature extractor based on the first hole convolutional neural network model to obtain the power grid state time series association feature graph;

[0022] A battery state multi-parameter time series association encoding unit is used to arrange the time series of the battery state into a battery state parameter time series joint matrix according to the parameter sample dimension and the time dimension, and then input the battery state parameter time series joint matrix into a battery state feature extractor based on a second void convolutional neural network model to obtain the battery state time series association feature graph.

[0023] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the characteristic autocorrelation strengthening module includes:

[0024] A self-attention feature enhancement unit, configured to input the grid state time series correlation feature graph and the battery state time series correlation feature graph into a cross-channel adaptive feature space structure enhanced self-attention module to obtain an enhanced grid state time series correlation feature graph and an enhanced battery state time series correlation feature graph;

[0025] The feature simplification and compression unit is used to perform global mean pooling processing on the enhanced power grid state timing correlation feature graph and the enhanced battery state timing correlation feature graph respectively to obtain the enhanced power grid state timing correlation feature vector and the enhanced battery state timing correlation feature vector.

[0026] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the self-attention feature enhancement unit includes:

[0027] A layer normalization subunit, used for performing layer normalization on the power grid state time series correlation feature graph to obtain a normalized power grid state time series correlation feature graph;

[0028] A channel context association encoding subunit, used for performing point convolution processing on the normalized power grid state time series association feature map to obtain a power grid state channel context association representation feature map;

[0029] A spatial context association encoding subunit, used for performing convolution encoding on the power grid state channel context association representation feature map to obtain a power grid state spatial context association representation feature map;

[0030] The global interactive feature fusion subunit is used to perform channel-space global interactive attention fusion on the power grid state channel context association representation feature map and the power grid state space context association representation feature map to obtain the enhanced power grid state time series association feature map.

[0031] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the global interactive feature fusion subunit is used to:

[0032] Copying the power grid state space context association representation characteristic graph to obtain a backup power grid state space context association representation characteristic graph;

[0033] Reshaping the grid state channel context association representation feature graph, the grid state space context association representation feature graph and the backup grid state space context association representation feature graph to obtain a grid state channel context association representation feature matrix, a grid state space context association representation feature matrix and a backup grid state space context association representation feature matrix;

[0034] Calculating a cross-channel cross covariance matrix between the power grid state channel context association representation feature matrix and the power grid state space context association representation feature matrix;

[0035] Using a Softmax function to activate the cross-channel cross-covariance matrix to obtain a global interactive attention matrix of power grid state features;

[0036] Calculating the product of the backup power grid state space context association representation feature matrix and the power grid state feature global interactive attention matrix to obtain an attention-enhanced power grid state feature representation matrix;

[0037] The attention-enhanced power grid state feature representation matrix is ​​reshaped to obtain the enhanced power grid state time series correlation feature map.

[0038] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the charging and discharging strategy generation module includes:

[0039] A probability reasoning unit, used for performing probability reasoning on the enhanced grid state time series associated feature vector, the enhanced battery state time series associated feature vector and the electricity price time series associated implicit feature vector to obtain a charge and discharge strategy time series reasoning representation vector;

[0040] The charge and discharge strategy generating unit is used to input the charge and discharge strategy time series reasoning representation vector into the charge and discharge decision module based on the classifier to obtain the charge and discharge strategy.

[0041] For example, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, the probability reasoning unit is used to:

[0042] The enhanced grid state time series associated feature vector, the enhanced battery state time series associated feature vector and the electricity price time series associated implicit feature vector are input into a charging and discharging strategy time series reasoning module based on a Bayesian belief network to obtain the charging and discharging strategy time series reasoning representation vector.

[0043] According to the overload and imbalance automatic adjustment compensation charging and discharging energy storage system of the embodiment of the present application, it can realize dynamic adjustment of the energy storage system, effectively compensate for the overload and imbalance problems of the power grid, improve equipment utilization, and at the same time ensure the stable operation of the power system and the power service quality of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0045] Figure 1 It shows a schematic diagram of the structure of the overload imbalance automatic adjustment compensation charging and discharging energy storage system in the embodiment of the present application;

[0046] Figure 2 It shows a schematic diagram of the structure of a controller of an overload imbalance automatic adjustment compensation charging and discharging energy storage system in an embodiment of the present application;

[0047] Figure 3 A flow chart of an overload imbalance automatic adjustment compensation charging and discharging energy storage method in an embodiment of the present application is shown;

[0048] Figure 4 A block diagram of another example of an overload imbalance automatic adjustment compensation charging and discharging energy storage system in an embodiment of the present application is shown;

[0049] Figure 5 A schematic diagram of another example of an overload imbalance automatic adjustment compensation charging and discharging energy storage system in an embodiment of the present application is shown;

[0050] Figure 6 A three-dimensional schematic diagram of a liquid-cooled battery box in an embodiment of the present application is shown;

[0051] Figure 7 A three-dimensional schematic diagram of a DC control box in an embodiment of the present application is shown;

[0052] Figure 8 A three-dimensional schematic diagram of an AC control box in an embodiment of the present application is shown; and

[0053] Fig. 9 A three-dimensional schematic diagram of an energy storage bidirectional converter in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0054] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0055] The terms used in this specification are those common terms currently widely used in the art in consideration of the functions of the present application, but these terms may vary according to the intentions of those skilled in the art, precedents, or new technologies in the art. In addition, specific terms may be selected, and in this case, their detailed meanings will be described in the detailed description of the present application. Therefore, the terms used in the specification should not be understood as simple names, but rather as a general description based on the meaning of the terms and the present application.

[0056] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0057] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. At the same time, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0058] In addition, the application architecture diagram in the embodiment of the present application is intended to more clearly illustrate the technical solution in the embodiment of the present application, and does not constitute a limitation on the technical solution provided in the embodiment of the present application. Of course, for other application architectures and business applications, the technical solution provided in the embodiment of the present application is also applicable to similar problems.

[0059] The following is a non-restrictive description of the overload imbalance automatic adjustment compensation charging and discharging energy storage system provided according to at least one embodiment of the present application through several examples or embodiments. As described below, different features in these specific examples or embodiments can be combined with each other without conflicting with each other to obtain new examples or embodiments, and these new examples or embodiments also belong to the scope of protection of this application.

[0060] In response to the above technical problems, the technical concept of this application is to use artificial intelligence technology based on deep learning to monitor and analyze the state of the power grid, electricity price information and energy storage battery status, to mine the time series change information of the power grid state, electricity price information and battery status, and then based on the dependency relationship between the three, intelligently generate charging and discharging strategies. In this way, dynamic adjustment of the energy storage system can be achieved, effectively compensating for the overload and imbalance problems of the power grid, improving equipment utilization, and ensuring the stable operation of the power system and the quality of power service for users.

[0061] Based on this, the present application proposes an overload imbalance automatic adjustment compensation charging and discharging energy storage system 1000, such as Figure 1 As shown, it includes: an input / output port 100, which is used to electrically connect to the power grid so as to absorb electric energy from the power grid or release electric energy to the power grid through the input / output port 100; a power conversion system (PCS) 200, which is used to convert the AC power of the power grid into DC power for storage or convert the stored DC power back into the AC power and supply it to the power grid or load; an energy storage battery 300, which is used to store and release electric energy; a battery management system 400, which is used to monitor and manage the charging and discharging process of the energy storage battery 300; and a controller 500, which is used to determine the charging and discharging strategy based on the power grid status, electricity price information and battery status.

[0062] Among them, Figure 2 As shown, the controller 500 includes: a power grid information acquisition module 510, which is used to acquire the time series of the electricity price information, the time series of the power grid state and the time series of the battery state, wherein the power grid state includes a voltage value, a current value and a frequency value, and the battery state includes a remaining power, a battery voltage value, a battery current value and a SOC; an electricity price information time series encoding module 520, which is used to perform time series encoding on the time series of the electricity price information to obtain an electricity price time series associated implicit feature vector; a multi-parameter time series association encoding module 530, which is used to perform state parameter time series association on the time series of the power grid state and the time series of the battery state, respectively. Encoding to obtain a grid state time series correlation feature diagram and a battery state time series correlation feature diagram; a feature autocorrelation reinforcement module 540, used to perform autocorrelation attention reinforcement on the grid state time series correlation feature diagram and the battery state time series correlation feature diagram respectively to obtain a reinforced grid state time series correlation feature vector and a reinforced battery state time series correlation feature vector; a charge and discharge strategy generation module 550, used to determine the charge and discharge strategy based on the probabilistic dependency relationship between the reinforced grid state time series correlation feature vector, the reinforced battery state time series correlation feature vector and the electricity price time series correlation implicit feature vector, wherein the charge and discharge strategy is used to indicate starting charging or starting discharging.

[0063] Specifically, the controller 500 first obtains the time series of electricity price information, the time series of grid state and the time series of battery state, wherein the grid state includes voltage value, current value and frequency value, and the battery state includes remaining power, battery voltage value, battery current value and SOC. It should be understood that electricity prices are usually affected by market supply and demand and energy prices and change in real time, which is an important factor affecting power operation and user electricity consumption behavior. By obtaining electricity price information in real time, the charging and discharging behavior of the energy storage system can be adjusted more accurately to obtain the best economic benefits. At the same time, grid state data such as voltage value, current value and frequency value reflect the real-time operation status of the grid, which helps to timely discover and understand the overload and imbalance problems of the grid. Battery state parameters such as remaining power, battery voltage value, battery current value and SOC (State of Charge) reflect the health status and remaining energy of the energy storage battery, which is the key basis for formulating charging and discharging strategies. In the technical solution of this application, by comprehensively analyzing electricity price information, grid state and battery state, it is helpful to realize intelligent decision-making of the energy storage system, ensuring that the economic benefits are maximized while meeting the grid stability and equipment life.

[0064] Next, in order to obtain the time series fluctuation information of electricity prices, the present application adopts an RNN-based sequence encoder to perform time series encoding on the time series of the electricity price information, so as to utilize the recurrent unit inside the RNN model to capture the dynamic change trend of electricity prices, reveal its potential periodic and seasonal change laws, and generate implicit feature vectors associated with the electricity price time series.

[0065] Correspondingly, the electricity price information time series encoding module 520 is used to: input the time series of the electricity price information into a sequence encoder based on RNN to obtain the electricity price time series associated implicit feature vector.

[0066] At the same time, considering that the grid state data and the battery state data respectively contain multiple state parameters, and each state parameter jointly describes the operational stability of the grid and the availability of the energy storage battery. Therefore, in order to more comprehensively understand the multi-dimensional characteristics of the grid state and the battery state, in the technical solution of the present application, first, the time series of the grid state and the time series of the battery state are arranged into a grid state parameter time series joint matrix and a battery state parameter time series joint matrix according to the parameter sample dimension and the time dimension, respectively, to integrate the multi-dimensional information of the grid state and the battery state, and maintain the continuity of each parameter over time. Then, different hole convolution neural network models are used to extract features from the grid state parameter time series joint matrix and the battery state parameter time series joint matrix, so as to utilize the hole convolution structure of the network model, while reducing the computational complexity, more effectively capture the spatiotemporal correlation relationship in the parameter time series joint matrix, learn the dynamic interaction mode between different state parameters and the time series evolution characteristics of the parameters, and generate a grid state time series correlation feature map and a battery state time series correlation feature map.

[0067] Correspondingly, the multi-parameter time series association coding module 530 includes: a power grid state multi-parameter time series association coding unit, which is used to arrange the time series of the power grid state into a power grid state parameter time series joint matrix according to the parameter sample dimension and the time dimension, and then input the power grid state parameter time series joint matrix into a power grid state feature extractor based on the first void convolutional neural network model to obtain the power grid state time series association feature graph; a battery state multi-parameter time series association coding unit, which is used to arrange the time series of the battery state into a battery state parameter time series joint matrix according to the parameter sample dimension and the time dimension, and then input the battery state parameter time series joint matrix into a battery state feature extractor based on the second void convolutional neural network model to obtain the battery state time series association feature graph.

[0068] Secondly, since the correlation between the various state parameters in the power grid state data and the battery state data is complex and dynamic, in the technical solution of the present application, in order to enhance the distinguishability and expressiveness of the feature representation, a cross-channel adaptive feature space structure enhanced self-attention module is further introduced to optimize and enhance the intrinsic feature structure of the power grid state time series correlation feature map and the battery state time series correlation feature map. Specifically, the cross-channel adaptive feature space structure enhanced self-attention module first performs layer normalization on the input feature map to eliminate the scale difference between layers and ensure the stability of feature processing. Then, the channel context association and spatial structure information of the feature map are captured through multi-layer convolution operations, and the channel context association information is used as a query, and the spatial structure information is used as a key and value. Based on the self-attention mechanism, the feature cross-channel global correlation interaction is performed. While retaining the spatial structure of the feature map, the channel correlation information between the feature structures is fused in units of the feature space structure, so as to enhance the feature interaction learning ability between different state parameters, and effectively reveal the nonlinear correlation relationship and potential spatiotemporal structure between the state parameters in the power grid state data and the battery state data, so as to obtain an enhanced power grid state time series correlation feature map and an enhanced battery state time series correlation feature map.

[0069] Correspondingly, the feature autocorrelation enhancement module 540 includes: a self-attention feature enhancement unit, which is used to input the power grid state timing correlation feature map and the battery state timing correlation feature map into the cross-channel adaptive feature space structure enhancement self-attention module to obtain an enhanced power grid state timing correlation feature map and an enhanced battery state timing correlation feature map; a feature simplification and compression unit, which is used to perform global mean pooling processing on the enhanced power grid state timing correlation feature map and the enhanced battery state timing correlation feature map respectively to obtain the enhanced power grid state timing correlation feature vector and the enhanced battery state timing correlation feature vector.

[0070] Among them, the self-attention feature enhancement unit includes: a layer normalization subunit, which is used to perform layer normalization on the power grid state time series association feature map to obtain a normalized power grid state time series association feature map; a channel context association encoding subunit, which is used to perform point convolution processing on the normalized power grid state time series association feature map to obtain a power grid state channel context association representation feature map; a spatial context association encoding subunit, which is used to perform convolution encoding on the power grid state channel context association representation feature map to obtain a power grid state spatial context association representation feature map; a global interactive feature fusion subunit, which is used to perform channel-space global interactive attention fusion on the power grid state channel context association representation feature map and the power grid state spatial context association representation feature map to obtain the enhanced power grid state time series association feature map.

[0071] Specifically, the global interactive feature fusion subunit is used to: copy the grid state space context association representation feature map to obtain a backup grid state space context association representation feature map; reshape the grid state channel context association representation feature map, the grid state space context association representation feature map and the backup grid state space context association representation feature map to obtain a grid state channel context association representation feature matrix, a grid state space context association representation feature matrix and a backup grid state space context association representation feature matrix; calculate the cross-channel cross covariance matrix between the grid state channel context association representation feature matrix and the grid state space context association representation feature matrix; activate the cross-channel cross covariance matrix using a Softmax function to obtain a grid state feature global interactive attention matrix; calculate the product between the backup grid state space context association representation feature matrix and the grid state feature global interactive attention matrix to obtain an attention-enhanced grid state feature representation matrix; and reshape the attention-enhanced grid state feature representation matrix to obtain the enhanced grid state time series association feature map.

[0072] In a specific example, the self-attention feature enhancement unit is used to: process the power grid state time series correlation feature graph using the following self-correlation attention enhancement formula to obtain the enhanced power grid state time series correlation feature graph, wherein the self-correlation attention enhancement formula is: in, represents the time series correlation characteristic diagram of the power grid state, Representation layer normalization operation, represents the normalized grid state time series correlation characteristic diagram, represents point convolution, A context-related representation feature graph representing the power grid state channel, Represents convolution processing based on 3×3 convolution kernel, represents the grid state space context association representation feature graph, Indicates a copy operation. represents the backup power grid state space context association representation feature graph, represents the feature shape reshaping, and They represent the power grid state channel context association representation feature matrix, the power grid state space context association representation feature matrix and the backup power grid state space context association representation feature matrix respectively. represents the cross-channel cross covariance matrix, is the scaling factor, represents the normalized exponential function, represents the matrix multiplication operation, A time series correlation characteristic diagram of the enhanced power grid state is shown.

[0073] Next, in order to improve the efficiency of feature representation and reduce the risk of overfitting, in the technical solution of the present application, the enhanced power grid state timing association feature graph and the enhanced battery state timing association feature graph are further subjected to global information compression through a global mean pooling operation to extract the main information of the feature graph, remove local details, and capture the global change trend of the state parameters to obtain an enhanced power grid state timing association feature vector and an enhanced battery state timing association feature vector.

[0074] Then, in order to comprehensively consider the actual operating status of the power grid, the charging and discharging capabilities of the energy storage battery, and the fluctuations in electricity prices, so as to more reasonably formulate the charging and discharging strategy of the energy storage system, the present application further adopts a charging and discharging strategy timing reasoning module based on the Bayesian belief network to perform probabilistic reasoning on the enhanced power grid state timing association feature vector, the enhanced battery state timing association feature vector, and the electricity price timing association implicit feature vector. The three are used as input nodes of the Bayesian belief network (BBN), and the Bayesian model is used to construct the conditional dependency relationship between the power grid state, the battery state and the electricity price. At the same time, the reinforcement learning algorithm is used to update the conditional probability table of each node in the BBN, fully considering the impact of the power grid state, the battery state and the electricity price on the charging and discharging strategy, and then the optimal charging and discharging decision probability distribution under various constraints is calculated to generate the optimal charging and discharging decision feature representation under current conditions, that is, the charging and discharging strategy timing reasoning representation vector.

[0075] Then, the charge and discharge strategy time-series reasoning representation vector is input into the classifier-based charge and discharge decision module to obtain the charge and discharge strategy, which is used to indicate starting charging or starting discharging. In the technical solution of the present application, the classifier-based charge and discharge decision module adopts a deep neural network structure, and performs feature learning and classification mapping on the charge and discharge strategy time-series reasoning representation vector through a fully connected network to determine whether the charging mode or the discharging mode should be started under the current grid status, battery status and electricity price conditions. In this way, real-time, dynamic and intelligent charge and discharge control of the energy storage system can be achieved, thereby maximizing economic benefits while ensuring grid stability, equipment life and battery health.

[0076] Correspondingly, the charging and discharging strategy generation module 550 includes: a probabilistic reasoning unit, which is used to perform probabilistic reasoning on the enhanced grid state time series association feature vector, the enhanced battery state time series association feature vector and the electricity price time series association implicit feature vector to obtain a charging and discharging strategy time series reasoning representation vector; and a charging and discharging strategy generation unit, which is used to input the charging and discharging strategy time series reasoning representation vector into a classifier-based charging and discharging decision module to obtain the charging and discharging strategy.

[0077] Among them, the probabilistic reasoning unit is used to: input the enhanced grid state timing association feature vector, the enhanced battery state timing association feature vector and the electricity price timing association implicit feature vector into the charging and discharging strategy timing reasoning module based on the Bayesian belief network to obtain the charging and discharging strategy timing reasoning representation vector.

[0078] It should be understood that the role of the classifier is to use the given categories and known training data to learn classification rules and classifiers, and then classify (or predict) unknown data. Logistic regression, SVM, etc. are often used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are required to form a multi-classification, but this is prone to errors and inefficient. Commonly used multi-classification methods include the Softmax classification function.

[0079] In a preferred example, it is considered that the enhanced grid state time series association feature vector, the enhanced battery state time series association feature vector and the electricity price time series association implicit feature vector respectively represent the time series implicit strengthening association feature between grid state parameters, the time series implicit strengthening association feature between battery state parameters and the time series implicit association feature of electricity price. Therefore, after the enhanced grid state time series association feature vector, the enhanced battery state time series association feature vector and the electricity price time series association implicit feature vector are input into the Bayesian belief network, the obtained charging and discharging strategy time series reasoning representation vector will also have insufficient coverage of fine-grained class probability reasoning due to the characteristic mode difference of the enhanced grid state time series association feature vector, the enhanced battery state time series association feature vector and the electricity price time series association implicit feature vector, thereby causing the outlier class regression reasoning mapping deviation, affecting the accuracy of the charging and discharging strategy obtained by the charging and discharging decision module based on the classifier input of the charging and discharging strategy time series reasoning representation vector.

[0080] Based on this, when the charge-discharge strategy temporal reasoning representation vector is input into the classifier-based charge-discharge decision module to obtain the charge-discharge strategy, the charge-discharge strategy temporal reasoning representation vector is optimized.

[0081] Correspondingly, the charge and discharge strategy generation unit includes: an optimization subunit, used to optimize the charge and discharge strategy timing reasoning representation vector to obtain an optimized charge and discharge strategy timing reasoning representation vector; a classification subunit, used to input the optimized charge and discharge strategy timing reasoning representation vector into the classifier-based charge and discharge decision module to obtain the charge and discharge strategy.

[0082] Specifically, the optimization subunit is used to: calculate the characteristic mean of the charge and discharge strategy timing reasoning representation vector, and divide the characteristic mean by the difference between the maximum eigenvalue and the minimum eigenvalue of the charge and discharge strategy timing reasoning representation vector to obtain the charge and discharge strategy timing reasoning representation distribution representation value; subtract the charge and discharge strategy timing reasoning representation distribution representation value and divide it by the charge and discharge strategy timing reasoning representation distribution representation value to obtain the charge and discharge strategy timing reasoning representation distribution modulation value; activate the charge and discharge strategy timing reasoning representation vector through a probabilistic function to obtain a probabilistic charge and discharge strategy timing reasoning representation vector; perform point subtraction between the probabilistic charge and discharge strategy timing reasoning representation vector and the charge and discharge strategy timing reasoning representation distribution modulation value, and take the absolute value. The value is calculated and the negative of the logarithmic value with base 2 is obtained to obtain the probabilistic charge and discharge strategy timing reasoning representation distribution modulation information vector; the distribution representation value of the charge and discharge strategy timing reasoning representation is divided by one minus the difference of each eigenvalue of the probabilistic charge and discharge strategy timing reasoning representation vector, and then all eigenvalues ​​of the probabilistic charge and discharge strategy timing reasoning representation vector are summed and divided by the length of the charge and discharge strategy timing reasoning representation vector to obtain the probabilistic charge and discharge strategy timing reasoning representation distribution modulation bias value; the probabilistic charge and discharge strategy timing reasoning representation distribution modulation information vector and the probabilistic charge and discharge strategy timing reasoning representation distribution modulation bias value and the weight as a hyperparameter are dotted to obtain the optimized charge and discharge strategy timing reasoning representation vector.

[0083] The optimization process of the timing reasoning representation vector of the charging and discharging strategy is expressed as follows: And among them: in, represents the characteristic mean of the timing reasoning representation vector of the charging and discharging strategy, and They respectively represent the maximum eigenvalue and the minimum eigenvalue in the timing reasoning representation vector of the charging and discharging strategy, represents the distribution characterization value of the timing reasoning of the charging and discharging strategy, represents the probabilistic charge-discharge strategy timing reasoning representation vector obtained after the charge-discharge strategy timing reasoning representation vector is activated by a probabilistic function, The first vector representing the probabilistic charging and discharging strategy timing reasoning is represented by eigenvalues, represents the logarithmic function value with base 2, is the weight as a hyperparameter, The value of is the length of the probabilistic charging and discharging strategy timing reasoning representation vector, Represents the timing reasoning representation vector of the optimized charging and discharging strategy.

[0084] That is, in the above preferred example, the eigenvalue-based probability information distribution planning of the charge-discharge strategy time series reasoning representation vector is performed through the Bernoulli probability modulation distribution of the charge-discharge strategy time series reasoning representation vector relative to the eigenvalue distribution, and the probability reverse mapping of the probability characteristics of the charge-discharge strategy time series reasoning representation vector as a whole is used as an extended coverage of the set mapping space of the charge-discharge strategy time series reasoning representation vector to independently understand the interaction path between the intuitive probability information distribution and the abstract probability space mapping of the charge-discharge strategy time series reasoning representation vector, so as to improve the accuracy of the charge-discharge strategy obtained by inputting the optimized charge-discharge strategy time series reasoning representation vector into the classifier-based charge-discharge decision module by avoiding the counterfactual reasoning mapping of the outlier feature distribution of the charge-discharge strategy time series reasoning representation vector to the class regression probability.

[0085] Based on the above embodiments, see Figure 3 As shown, it is a flow chart of an overload imbalance automatic adjustment compensation charging and discharging energy storage method in an embodiment of the present application. Figure 3 As shown, according to the overload imbalance automatic adjustment compensation charging and discharging energy storage method of the embodiment of the present application, the steps include: S510, obtaining the time series of electricity price information, the time series of power grid state and the time series of battery state, wherein the power grid state includes voltage value, current value and frequency value, and the battery state includes remaining power, battery voltage value, battery current value and SOC; S520, time series encoding the time series of the electricity price information to obtain the electricity price time series associated implicit feature vector; S530, respectively performing state parameter time series correlation on the time series of the power grid state and the time series of the battery state S540, respectively, performing autocorrelation attention enhancement on the grid state timing association feature graph and the battery state timing association feature graph to obtain an enhanced grid state timing association feature vector and an enhanced battery state timing association feature vector; S550, based on the probabilistic dependency relationship between the enhanced grid state timing association feature vector, the enhanced battery state timing association feature vector and the electricity price timing association implicit feature vector, determining the charge and discharge strategy, wherein the charge and discharge strategy is used to indicate starting charging or starting discharging.

[0086] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned overload imbalance automatic adjustment compensation charging and discharging energy storage method have been referred to above. Figure 2 The overload imbalance automatic adjustment compensation controller 500 in the charging and discharging energy storage system 1000 has been described in detail, and therefore, its repeated description will be omitted.

[0087] In another example of the present application, an overload imbalance automatic adjustment compensation charging and discharging micro energy storage system is also provided, referring to Figure 4 and Figure 5 As shown, it can automatically adjust the charging and discharging time according to the nature of the power load, automatically charge and discharge the three-phase imbalance, control valley charging and peak discharge, realize PCS inverter single and three-phase output, modular design of installed capacity, can increase and decrease battery capacity and power size at any time, quick plug-in design, and no wiring function. The BMS battery management system is adopted to realize the comprehensive thermal management and fire extinguishing device of LiFePO4 lithium battery. Liquid cooling or air cooling device can be selected according to the actual environmental conditions. In addition, the current and voltage conditions can be collected in real time through the EMS energy management cloud platform, and intelligently uploaded to realize the mobile phone APP function monitoring and alarm functions.

[0088] Among them, the power conversion system (PCS), also known as the energy storage inverter, is the core power conversion unit in the energy storage device. It adopts a three-level topology structure and can realize bidirectional conversion from DC to AC and AC to DC. It can convert AC power into DC power to charge the battery, and can also convert DC power into AC power to power the load or feed it back to the grid.

[0089] The power conversion system (PCS) adopts a three-level topology, with a maximum conversion efficiency of >98.5%, high dynamic response, full-load switching time as low as 10ms, supports multi-machine parallel operation, and can be expanded to 2MW. Modular equipment makes configuration more flexible and maintenance convenient. The use of high-speed IGBT and low internal resistance filter can achieve seamless switching between on-grid and off-grid, fast response, and uninterrupted power supply.

[0090] Further, Figure 6 The three-dimensional schematic diagram of the liquid-cooled battery box is shown. The liquid-cooled battery box is provided with multiple groups such as Figure 5 The multiple battery modules shown in the figure can be cooled by liquid cooling. Through the liquid cooling design, the temperature difference between the multiple battery modules is less than 2°C. Figure 7 Shows Figure 5The DC control box includes the main control BMU and pre-charge EMCU. It has functions such as fault alarm, fault protection, and safety protection. It also collects information such as air conditioning, fire protection, and electric meters, and interacts with EMS and PCS through Ethernet. It responds to the control strategy of EMS and realizes energy scheduling of the energy storage system, system thermal management, and fault handling. Figure 8 Shows Figure 5 The three-dimensional schematic diagram of the AC control box in the figure. The AC control box can control multiple AC circuits. It is equipped with control devices, fuses and obvious power-off devices. It is flexible to install and use. It has functions such as overcurrent, overheating and tripping protection, which can ensure that the primary and secondary circuits are cut off in time. Fig. 9 Shows Figure 5 The three-dimensional schematic diagram of the energy storage bidirectional converter in the figure. The energy storage bidirectional converter can convert alternating current (AC) into direct current (DC), and can also convert direct current (DC) into alternating current (AC), and can perform bidirectional conversion.

[0091] Based on the above embodiments, another exemplary embodiment of an electronic device is also provided in the embodiments of the present application. In some possible implementations, the electronic device in the embodiments of the present application may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement the steps of the overload imbalance automatic adjustment compensation charging and discharging energy storage method in the above embodiments when executing the program.

[0092] The embodiment of the present application also provides a computer-readable storage medium, on which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, the overload imbalance automatic adjustment compensation charging and discharging energy storage method according to the embodiment of the present application described with reference to the above figures can be executed. The computer-readable storage medium includes, but is not limited to, for example, volatile memory and / or non-volatile memory. The volatile memory may, for example, include a random access memory (RAM) and / or a cache memory (cache), etc. The non-volatile memory may, for example, include a read-only memory (ROM), a hard disk, a flash memory, etc.

[0093] The embodiment of the present application also provides a computer program product or a computer program, which includes computer executable instructions, and the computer executable instructions are stored in a computer readable storage medium. The processor of the computer device reads the computer executable instructions from the computer readable storage medium, and the processor executes the computer executable instructions, so that the computer device executes the overload imbalance automatic adjustment compensation charging and discharging energy storage method according to the embodiment of the present application.

[0094] Those skilled in the art will appreciate that the contents disclosed in this application may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

[0095] In addition, although the present application makes various references to certain units in the system according to embodiments of the present application, any number of different units can be used and run on the client and / or server. The units are only illustrative, and different aspects of the system and method can use different units.

[0096] Those skilled in the art will appreciate that all or part of the steps in the above method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk or an optical disk. Optionally, all or part of the steps in the above embodiment can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiment can be implemented in the form of hardware or in the form of software function modules. The present application is not limited to any particular form of combination of hardware and software.

[0097] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those of ordinary skill in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or extremely formal sense, unless explicitly defined as such herein.

[0098] The above is an explanation of the present application and should not be considered as limiting thereof. Although several exemplary embodiments of the present application have been described, those skilled in the art will readily appreciate that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present application.

Claims

1. An overload imbalance automatic adjustment compensation charging and discharging energy storage system, characterized in that: include: An input / output port, the input / output port being used to be electrically connected to a power grid so as to absorb electric energy from the power grid or release electric energy to the power grid through the input / output port; A power conversion system PCS, used to convert the AC power of the power grid into DC power for storage or to convert the stored DC power back into AC power and supply it to the power grid or load; Energy storage batteries, used to store and release electrical energy; A battery management system, used to monitor and manage the charging and discharging process of the energy storage battery; A controller for determining a charging and discharging strategy based on a grid status, electricity price information, and a battery status; The controller comprises: A power grid information acquisition module, used to acquire the time series of the electricity price information, the time series of the power grid state and the time series of the battery state, wherein the power grid state includes a voltage value, a current value and a frequency value, and the battery state includes a remaining power, a battery voltage value, a battery current value and a SOC; An electricity price information time series encoding module, used for performing time series encoding on the time series of the electricity price information to obtain an electricity price time series associated implicit feature vector; A multi-parameter time series correlation coding module, used for performing time series correlation coding of state parameters on the time series of the power grid state and the time series of the battery state to obtain a time series correlation characteristic diagram of the power grid state and a time series correlation characteristic diagram of the battery state; A feature autocorrelation reinforcement module, used to perform autocorrelation attention reinforcement on the grid state time series correlation feature graph and the battery state time series correlation feature graph respectively to obtain a reinforced grid state time series correlation feature vector and a reinforced battery state time series correlation feature vector; A charge and discharge strategy generation module, used to determine the charge and discharge strategy based on the probability dependency relationship between the enhanced grid state time series associated feature vector, the enhanced battery state time series associated feature vector and the electricity price time series associated implicit feature vector, wherein the charge and discharge strategy is used to indicate starting charging or starting discharging; The charging and discharging strategy generation module includes: A probability reasoning unit, used for performing probability reasoning on the enhanced grid state time series associated feature vector, the enhanced battery state time series associated feature vector and the electricity price time series associated implicit feature vector to obtain a charge and discharge strategy time series reasoning representation vector; A charge-discharge strategy generating unit, used for inputting the charge-discharge strategy time series reasoning representation vector into a charge-discharge decision module based on a classifier to obtain the charge-discharge strategy; The charge and discharge strategy generation unit includes: an optimization subunit, which is used to optimize the charge and discharge strategy time series reasoning representation vector to obtain an optimized charge and discharge strategy time series reasoning representation vector; a classification subunit, which is used to input the optimized charge and discharge strategy time series reasoning representation vector into the classifier-based charge and discharge decision module to obtain the charge and discharge strategy; The optimization subunit is used to: calculate the characteristic mean of the charge and discharge strategy timing reasoning representation vector, and divide the characteristic mean by the difference between the maximum eigenvalue and the minimum eigenvalue of the charge and discharge strategy timing reasoning representation vector to obtain the charge and discharge strategy timing reasoning representation distribution representation value; subtract the charge and discharge strategy timing reasoning representation distribution representation value and divide it by the charge and discharge strategy timing reasoning representation distribution representation value to obtain the charge and discharge strategy timing reasoning representation distribution modulation value; activate the charge and discharge strategy timing reasoning representation vector through a probabilistic function to obtain a probabilistic charge and discharge strategy timing reasoning representation vector; perform point subtraction between the probabilistic charge and discharge strategy timing reasoning representation vector and the charge and discharge strategy timing reasoning representation distribution modulation value, and take the absolute value And calculate the negative of the logarithmic value with base 2 to obtain the probabilistic charge and discharge strategy timing reasoning representation distribution modulation information vector; divide the distribution representation value of the charge and discharge strategy timing reasoning representation by one minus the difference of each eigenvalue of the probabilistic charge and discharge strategy timing reasoning representation vector, sum all eigenvalues ​​of the probabilistic charge and discharge strategy timing reasoning representation vector, and divide it by the length of the charge and discharge strategy timing reasoning representation vector to obtain the probabilistic charge and discharge strategy timing reasoning representation distribution modulation bias value; perform dot addition on the probabilistic charge and discharge strategy timing reasoning representation distribution modulation information vector and the probabilistic charge and discharge strategy timing reasoning representation distribution modulation bias value and the weight as a hyperparameter to obtain the optimized charge and discharge strategy timing reasoning representation vector.

2. The overload imbalance automatic adjustment compensation charging and discharging energy storage system according to claim 1 is characterized in that: The electricity price information timing coding module is used to: The time series of the electricity price information is input into a sequence encoder based on RNN to obtain the electricity price time series associated implicit feature vector.

3. The overload imbalance automatic adjustment compensation charging and discharging energy storage system according to claim 2 is characterized in that: The multi-parameter temporal association encoding module includes: A power grid state multi-parameter time series association encoding unit, used for arranging the time series of the power grid state into a power grid state parameter time series joint matrix according to the parameter sample dimension and the time dimension, and inputting the power grid state parameter time series joint matrix into a power grid state feature extractor based on the first hole convolutional neural network model to obtain the power grid state time series association feature graph; A battery state multi-parameter time series association encoding unit is used to arrange the time series of the battery state into a battery state parameter time series joint matrix according to the parameter sample dimension and the time dimension, and then input the battery state parameter time series joint matrix into a battery state feature extractor based on a second void convolutional neural network model to obtain the battery state time series association feature graph.

4. The overload imbalance automatic adjustment compensation charging and discharging energy storage system according to claim 3 is characterized in that: The feature autocorrelation enhancement module includes: A self-attention feature enhancement unit, configured to input the grid state time series correlation feature graph and the battery state time series correlation feature graph into a cross-channel adaptive feature space structure enhanced self-attention module to obtain an enhanced grid state time series correlation feature graph and an enhanced battery state time series correlation feature graph; The feature simplification and compression unit is used to perform global mean pooling processing on the enhanced power grid state timing correlation feature graph and the enhanced battery state timing correlation feature graph respectively to obtain the enhanced power grid state timing correlation feature vector and the enhanced battery state timing correlation feature vector.

5. The overload imbalance automatic adjustment compensation charging and discharging energy storage system according to claim 4 is characterized in that: The self-attention feature enhancement unit comprises: A layer normalization subunit, used for performing layer normalization on the power grid state time series correlation feature graph to obtain a normalized power grid state time series correlation feature graph; A channel context association encoding subunit, used for performing point convolution processing on the normalized power grid state time series association feature map to obtain a power grid state channel context association representation feature map; A spatial context association encoding subunit, used for performing convolution encoding on the power grid state channel context association representation feature map to obtain a power grid state spatial context association representation feature map; The global interactive feature fusion subunit is used to perform channel-space global interactive attention fusion on the power grid state channel context association representation feature map and the power grid state space context association representation feature map to obtain the enhanced power grid state time series association feature map.

6. The overload imbalance automatic adjustment compensation charging and discharging energy storage system according to claim 5 is characterized in that: The global interactive feature fusion subunit is used to: Copying the power grid state space context association representation characteristic graph to obtain a backup power grid state space context association representation characteristic graph; Reshaping the grid state channel context association representation feature graph, the grid state space context association representation feature graph and the backup grid state space context association representation feature graph to obtain a grid state channel context association representation feature matrix, a grid state space context association representation feature matrix and a backup grid state space context association representation feature matrix; Calculating a cross-channel cross covariance matrix between the power grid state channel context association representation feature matrix and the power grid state space context association representation feature matrix; Using a Softmax function to activate the cross-channel cross-covariance matrix to obtain a global interactive attention matrix of power grid state features; Calculating the product of the backup power grid state space context association representation feature matrix and the power grid state feature global interactive attention matrix to obtain an attention-enhanced power grid state feature representation matrix; The attention-enhanced power grid state feature representation matrix is ​​reshaped to obtain the enhanced power grid state time series correlation feature map.

7. The overload imbalance automatic adjustment compensation charging and discharging energy storage system according to claim 6 is characterized in that: The probability reasoning unit is used to: The enhanced grid state time series associated feature vector, the enhanced battery state time series associated feature vector and the electricity price time series associated implicit feature vector are input into a charging and discharging strategy time series reasoning module based on a Bayesian belief network to obtain the charging and discharging strategy time series reasoning representation vector.

Images