Energy management system

Abstract

The invention relates to the technical field of energy management, and particularly discloses an energy management system which monitors the working states of battery modules by collecting state parameters of the battery modules in real time and introduces a deep learning algorithm to perform charging state time sequence analysis on the battery modules in a charging state. According to the method, the time sequence charging characteristics of all the battery modules are captured through the global time sequence charging characteristics of the battery modules, then charging state query analysis is conducted on all the battery modules on the basis of the global time sequence charging characteristics of the battery modules, and abnormal monitoring and charging strategy management of the charging states of all the battery modules are achieved. Through the mode, the charging state of each battery module can be obtained more effectively, potential charging abnormity can be found and processed in time, and the charging process is optimized, so that the efficiency and the reliability of the energy management system are improved.

Description

Technical Field

[0001] This application relates to the technical field of energy management, and more specifically, to an energy management system. Background Art

[0002] Energy storage EMS (ENERGY MANAGEMENT SYSTEM) is a set of software and hardware intelligent systems integrating monitoring, control, analysis, and optimization, specifically used for the management of energy systems. EMS controls the orderly and stable operation of the entire energy storage system through data collection, data analysis and display, and energy scheduling of energy storage system devices (PCS, BMS, electricity meters, fire protection, air conditioners, etc.).

[0003] As the core unit of energy storage, the performance stability and safety of the battery module are directly related to the efficiency and reliability of the entire system. However, during the use of the battery module, it will face various complex factors, such as temperature fluctuations, inconsistent charge and discharge rates, and different degrees of aging. These factors will all lead to a decline in the performance of the battery module and even pose potential safety hazards.

[0004] Traditional battery management methods mainly rely on static parameter monitoring and preset threshold judgment. For example, basic information such as battery voltage, current, and temperature is monitored, and fixed safety thresholds are set to prevent problems such as overcharging, over-discharging, and overheating. Although this method can ensure the basic safe operation of the battery module to a certain extent, due to the high dynamics and uncertainty of the performance and use environment of the battery module, traditional static parameter monitoring and preset threshold judgment methods are often difficult to adapt to various dynamic changes of the battery module during actual operation, which may lead to the battery management system being unable to identify and handle potential risks of the battery module in a timely and accurate manner, thereby affecting the overall performance and safety of the system.

[0005] Therefore, an optimized energy management system is expected. Summary of the Invention

[0006] In order to solve the above technical problems, this application is proposed. Embodiments of this application provide an energy management system, which monitors the working state of battery modules by real-time collecting the state parameters of each battery module, and introduces a deep learning algorithm to perform charging state time series analysis on each battery module in the charging state to capture the time series charging characteristics of each battery module. Furthermore, based on the time series charging characteristics of the global battery modules, charging state query analysis is performed on each battery module to achieve abnormal monitoring of the charging state of each battery module and charging strategy management. In this way, the charging state of each battery module can be obtained more effectively, potential charging anomalies can be discovered and processed in a timely manner, the charging process can be optimized, thereby improving the efficiency and reliability of the energy management system.

[0007] Accordingly, in one aspect of the present application, an energy management system is provided, which includes:

[0008] A data acquisition module, a status monitoring module, and an energy control module;

[0009] The data acquisition module is used to acquire the status parameters of each battery module, and the status parameters include voltage, current, and temperature;

[0010] The status monitoring module is used to perform status monitoring on each battery module based on the status parameters of each battery module to obtain a set of status monitoring results. Among them, performing status monitoring on each battery module includes taking each battery module in the charging state as the monitoring object, and performing status query analysis based on the charging characteristics of the global battery module to obtain the set of status monitoring results;

[0011] The energy control module is used to determine the charging strategy for the battery module based on the situation of the power grid.

[0012] Compared with the prior art, the energy management system provided by the present application monitors the working status of the battery module by real-time acquiring the status parameters of each battery module, and introduces a deep learning algorithm to perform charging state time series analysis on each battery module in the charging state to capture the time series charging characteristics of each battery module. Furthermore, based on the time series charging characteristics of the global battery module, status query analysis is performed on each battery module to achieve abnormal monitoring of the charging state of each battery module and charging strategy management. In this way, the charging state of each battery module can be obtained more effectively, potential charging anomalies can be discovered and processed in a timely manner, the charging process can be optimized, thereby improving the efficiency and reliability of the energy management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other objects, features, and advantages of the present application will become more obvious. The accompanying drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation to the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0014] Figure 1 It is a block diagram of an energy management system according to an embodiment of the present application.

[0015] Figure 2 It is a block diagram of a status monitoring module in an energy management system according to an embodiment of the present application.

[0016] Figure 3It is a schematic diagram of data flow in the status monitoring module in the energy management system according to an embodiment of the present application.

[0017] Figure 4 It is a block diagram of the charging status query unit in the energy management system according to an embodiment of the present application.

[0018] Figure 5 It is a block diagram of the charging status query response analysis subunit in the energy management system according to an embodiment of the present application. Detailed implementation manners

[0019] As shown in the present application and the claims, unless the context clearly indicates otherwise, words such as "a", "an", "one", and / or "the" are not specifically singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0020] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or the server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0021] 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 operations before or below do not necessarily need to be executed precisely in order. On the contrary, as needed, they can be executed in reverse order or processed simultaneously. At the same time, other operations can also be added to these processes, or one or several operations can be removed from these processes.

[0022] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0023] In view of the technical problems described in the above background art, the present application proposes an energy management system. It monitors the working status of battery modules by collecting the status parameters of each battery module in real time, and introduces a deep learning algorithm to perform a charging status time series analysis on each battery module in the charging state to capture the time series charging characteristics of each battery module. Furthermore, based on the time series charging characteristics of the global battery modules, it conducts a charging status query analysis on each battery module to achieve abnormal monitoring of the charging status of each battery module and charging strategy management. In this way, it can more effectively obtain the charging status of each battery module, timely detect and handle potential charging anomalies, optimize the charging process, thereby improving the efficiency and reliability of the energy management system.

[0024] Figure 1 FIG. is a block diagram of an energy management system according to an embodiment of the present application. As Figure 1 shown, the energy management system 100 includes: the data acquisition module 110, which is used to collect the status parameters of each battery module, and the status parameters include voltage, current, and temperature; the status monitoring module 120, which is used to perform status monitoring on each battery module based on the status parameters of each battery module to obtain a set of status monitoring results. Among them, performing status monitoring on each battery module includes taking each battery module in the charging state as the monitoring object and performing a status query analysis based on the charging characteristics of the global battery modules to obtain the set of status monitoring results; the energy control module 130, which is used to determine the charging strategy for the battery module based on the situation of the power grid.

[0025] In the above energy management system, the data acquisition module 110 is used to collect the status parameters of each battery module, and the status parameters include voltage, current, and temperature. It should be understood that the voltage can intuitively reflect the state of charge and internal electrochemical performance of the battery; the current is directly related to the charge and discharge rate and power output of the battery; and the temperature has an important impact on the chemical reaction rate, life, and safety of the battery. By collecting the status parameters of each battery module in real time, it is possible to achieve a comprehensive monitoring of the real-time status of the battery module, providing basic data for the abnormal status analysis and charging strategy management of the battery module.

[0026] Specifically, for the acquisition of voltage data, high-precision voltage sensors can be used. Such sensors need to have excellent anti-electromagnetic interference capabilities because strong electromagnetic fields that may exist in the actual working environment will seriously affect the measurement results. To achieve this, not only carefully selected sensors with strong anti-interference performance are required, but also shielded cables need to be used for connection. This kind of cable can effectively isolate the interference of external electromagnetic fields and ensure the purity of signal transmission. At the same time, considering the possible drift phenomenon of the sensor during long-term use, regular calibration is particularly important. Through periodic calibration, the working state of the sensor can be adjusted in time to ensure that it always remains within the optimal performance range, thereby providing stable and reliable voltage data.

[0027] The measurement of current mainly relies on Hall effect current sensors or shunts. Hall effect current sensors indirectly calculate the current magnitude by detecting the magnetic field change generated when current flows through, and this method does not need to be directly connected to the circuit, so it will not affect the original circuit. While the shunt is based on Ohm's law and determines the current intensity by measuring the voltage drop across an element with a known resistance value when current passes through. Each of the two methods has its own advantages and disadvantages, and factors such as cost, installation space, and required accuracy need to be comprehensively considered when making a choice. No matter which method is adopted, special attention needs to be paid to preventing the sensor from being damaged by the thermal effect caused by large currents, and effective heat dissipation measures need to be taken to maintain the stability of the sensor during operation. In addition, for the possible electromagnetic interference problems, shielding technology also needs to be adopted to ensure the accuracy and stability of the current measurement data.

[0028] As another key parameter, the measurement accuracy of temperature directly affects the safe operation of the battery. Traditional thermistors have been widely used in the temperature monitoring of individual battery cells due to their advantages such as fast response speed and small size. At the same time, an additional over-temperature protection mechanism can be set to trigger an alarm and take corresponding protective measures immediately once an abnormal high temperature is detected. In addition, considering that the performance of temperature sensors may vary under different environmental conditions, the sensors also need to be appropriately calibrated according to the specific application scenario to ensure the consistency and reliability of the measurement results.

[0029] In addition to hardware-level optimizations, software support is also an important part of the data acquisition process. For example, developing efficient data processing algorithms to filter out error signals caused by accidental factors and extract truly valuable information. These algorithms need to take into account various possible interference sources and their influence patterns, and analyze and correct them through mathematical models to improve the quality of data. Additionally, formulating a reasonable communication protocol is also crucial, which determines whether the data collected from numerous sensors can be transmitted to the central processing unit quickly and accurately. An efficient communication protocol can not only reduce the risk of data loss but also speed up the information update rate, making the decision-making process more agile. On this basis, an intelligent diagnosis function can be introduced to automatically identify potential problem trends through learning historical data and issue early warnings, further enhancing the proactive defense ability of the system.

[0030] Considering the reliability of the long-term operation of the data acquisition module, implementing the redundancy design principle cannot be ignored. For example, deploying backup sensors at critical positions. When the main sensor fails, the backup sensor can immediately take over the work to avoid data interruption caused by a single-point failure. At the same time, formulating a unified standard interface specification to promote the compatibility and interoperability between different devices, simplify the system integration process, reduce maintenance costs, and enhance the flexibility and scalability of the overall system.

[0031] In the above energy management system, the state monitoring module 120 is used to perform state monitoring on each battery module based on the state parameters of each battery module to obtain a set of state monitoring results. Among them, performing state monitoring on each battery module includes taking each battery module in the charging state as the monitoring object and performing state query analysis based on the charging characteristics of the global battery module to obtain the set of state monitoring results. Specifically, this application takes into account that the charging process of the battery module is the most critical link in its life cycle and also a high-incidence period of faults. Therefore, it is particularly important to monitor and manage the charging process of each battery module. Due to the electrochemical characteristics, manufacturing process differences, and diversity of the use environment inside the battery module, individual battery modules may experience abnormal conditions such as overcharging, over-discharging, and imbalance during the charging process. To finely manage the charging state of each battery module, this application takes each battery module in the charging state as the monitoring object, and through state query of each battery module based on the charging characteristics of the global battery module, deeply analyzes the charging behavior patterns of each battery module within the framework of the overall change law of the charging characteristics of the battery module, so as to identify the abnormalities in the charging state of individual battery modules from a combination of macroscopic and microscopic perspectives, thereby providing a key basis for timely adjusting the charging strategy and avoiding the deterioration of battery module performance and safety accidents in the future. Among them, Figure 2 It is a block diagram of the state monitoring module in the energy management system according to the embodiment of the present application.Figure 3 It is a schematic diagram of data flow in the state monitoring module in the energy management system according to an embodiment of the present application. As Figure 2 and Figure 3 shown, the state monitoring module 120 includes: a charging state battery module information acquisition unit 121, configured to extract the state parameters of the battery modules in the charging state from the state parameters of the battery modules to obtain a set of charging battery module state parameters, where the charging battery module state parameters include the time series of voltage, the time series of current, and the time series of temperature; a charging state time series analysis unit 122, configured to perform charging state time series correlation feature extraction on each charging battery module state parameter in the set of charging battery module state parameters to obtain a set of charging battery module charging state time series correlation feature coding vectors; a query feature extraction unit 123, configured to extract the charging battery module charging state time series correlation feature coding vector of the detected battery module from the set of charging battery module charging state time series correlation feature coding vectors as a charging state query feature vector; a charging state query unit 124, configured to determine the state monitoring result of the detected battery module based on the feature interaction response pattern of the charging state query feature vector with respect to the set of charging battery module charging state time series correlation feature coding vectors.

[0032] Specifically, the charging state battery module information acquisition unit 121 is configured to extract the state parameters of the battery modules in the charging state from the state parameters of the battery modules to obtain a set of charging battery module state parameters, where the charging battery module state parameters include the time series of voltage, the time series of current, and the time series of temperature. It should be understood that during the charging process, the voltage, current, and temperature of the battery modules change dynamically over time, and the parameter values at a single time point are difficult to comprehensively reflect the charging state of the battery modules. Therefore, in the present application, by extracting the time series data of voltage, current, and temperature of each battery module in the charging state, the dynamic change behavior of the battery modules during the charging process can be understood more accurately, and thus the abnormal state of the battery modules during the charging process can be captured and identified more accurately.

[0033] Specifically, which battery modules are in the charging state is usually clearly identified by monitoring the working mode flag bits of each battery module, and these flag bits are dynamically updated by the Battery Management System (BMS) according to the real-time working conditions. For the time series acquisition of voltage, high-precision voltage sensors are deployed across the selected charging battery modules to continuously record voltage changes and store the data in the form of timestamps. Since voltage fluctuations are relatively frequent during the charging process, the sensors are required to have fast response capabilities and high resolution to capture the slightest changes. In addition, to ensure data consistency during long-term operation, it is an essential step to calibrate the sensors regularly. In this way, an accurate trajectory of the voltage change over time during the charging cycle of each charging battery module can be obtained.

[0034] The time series acquisition of current relies on Hall effect current sensors or shunts. Which method to choose depends on the specific application scenario and requirements. Hall effect current sensors indirectly measure the current magnitude by detecting the magnetic field changes generated when current flows through a conductor. This method does not require direct connection to the circuit, avoiding the possible risks of interference or damage. A shunt, on the other hand, uses Ohm's law to calculate the current intensity by measuring the voltage drop across a component with a known resistance value when current passes through it. Whatever method is adopted, special attention should be paid to the thermal effect problems that may occur under high current conditions, and appropriate heat dissipation measures should be taken. At the same time, to ensure the continuity and accuracy of current data, the sensors also need to be equipped with a timestamp function so that each record can correspond to a specific moment. In this way, a solid foundation is established for analyzing the variation law of current over time during the charging process.

[0035] The time series acquisition of temperature involves the application of thermistors or digital temperature sensors. As the charging process progresses, the sensors continuously record temperature data and attach corresponding time marks. To ensure the authenticity and reliability of the data, in addition to selecting high-performance sensors, it is also necessary to consider how to effectively cope with the temperature difference effects under different environmental conditions. For example, for battery modules operating in a high-temperature environment, their temperature sensors may face greater challenges. Therefore, in practical applications, the position of the sensors should be adjusted according to the specific situation to ensure that any abnormal temperature rise can be accurately captured.

[0036] During the entire data acquisition process, synchronization is a factor that cannot be ignored. To ensure the time consistency among different parameters such as voltage, current, and temperature, a unified time reference needs to be established. This is usually achieved with the help of a high-precision clock module, which provides standard time signals for all sensors, enabling each record to be accurately mapped to the same time coordinate system. This facilitates the subsequent analysis phase to directly compare the change trends of various parameters based on the same time axis, making it easier to discover potential problems or optimization spaces. For example, when it is observed that the voltage rises sharply within a certain period while the corresponding current does not increase significantly, it can be inferred that there may be local overheating or other potential fault hazards, and then timely measures can be taken to address them.

[0037] Meanwhile, to ensure the stability and efficiency of data transmission, building an efficient and reliable communication network is also a crucial step. Whether it is wired or wireless communication, it is necessary to ensure low-latency and high-bandwidth data transmission performance under various complex working conditions. Especially in large-scale energy storage systems, numerous sensors are distributed at different physical locations, and how to coordinate data exchange between these devices becomes particularly important. Therefore, formulating reasonable communication protocols and routing strategies, and optimizing the data packet format and transmission path can greatly reduce the probability of information loss and repeated transmission, improving the overall system's response speed and processing capacity.

[0038] In addition, for possible data anomalies, corresponding fault tolerance mechanisms need to be set up. For example, when a certain sensor fails to report the latest data on time due to reasons, the system should be able to automatically identify and attempt to re-acquire the information, or temporarily use prediction algorithms to fill in the missing values to avoid affecting the accuracy of subsequent analysis results. At the same time, regularly checking and maintaining the hardware facilities and timely replacing aging or damaged components are also important guarantees for long-term stable operation. Through the above various measures, it is ensured that the state parameters of the rechargeable battery module such as voltage, current, and temperature are rich, detailed, and accurate, laying a good foundation for in-depth research on charging behavior characteristics and their impact on the battery's health status.

[0039] Specifically, the charging state timing analysis unit 122 is configured to extract charging state timing correlation feature encoding vectors of each charging battery module state parameter in the set of charging battery module state parameters, so as to obtain a set of charging battery module charging state timing correlation feature encoding vectors. In a specific example of the present application, a charging state timing encoder based on the LSTM model is used to perform timing correlation analysis on each charging battery module state parameter in the set of charging battery module state parameters, so as to obtain the set of charging battery module charging state timing correlation feature encoding vectors. It should be understood that during the charging process of the battery module, there are inherent timing correlations among state parameters such as voltage, current, and temperature. For example, the increase in voltage is often accompanied by the inflow of current, and the increase in temperature may be the result of an accelerated chemical reaction rate. Therefore, in order to capture the timing correlation change patterns among various state parameters, the present application uses the LSTM model to construct a charging state timing encoder, and performs timing correlation analysis on the state parameter data of each charging battery module respectively, so as to capture the timing correlation change characteristics among multi-source state parameters during the charging process of each charging battery module. Those of ordinary skill in the art should know that the LSTM model (long short-term memory network) internally includes an input gate, a forget gate, and an output gate. Through the collaborative work of the gate structures, it can effectively control the flow and memory of information in the time series, so as to realize the capture and modeling of long-term dependence relationships in time series data. When processing the time series data of the voltage, current, and temperature of the charging battery module, the voltage, current, and temperature data of each time step are sequentially input into the LSTM model. Through the selective memory and forgetting mechanism of the forget gate and the input gate, the input data of each time step is combined with the hidden state of the previous moment to update the hidden state of the current moment, so as to capture the dynamic correlation and timing change trends among parameters such as voltage, current, and temperature during the charging of the battery module, realize the learning and modeling of the timing correlation characteristics of multi-source state parameters during the charging process of the battery module, and finally output an encoding vector representation that can characterize the timing correlation change characteristics of each charging battery module charging state, so as to obtain a set of charging battery module charging state timing correlation feature encoding vectors.

[0040] Specifically, the query feature extraction unit 123 is configured to extract the charging battery module charging state timing correlation feature encoding vector of the detected battery module from the set of charging battery module charging state timing correlation feature encoding vectors as the charging state query feature vector. It should be understood that due to factors such as manufacturing process, use environment, and aging degree, the charging characteristics of each battery module are different. Therefore, by extracting the charging battery module charging state timing correlation feature encoding vector corresponding to the detected battery module as the charging state query feature vector, the present application helps to focus on the personalized charging behavior characteristics of this module and perform targeted charging anomaly analysis.

[0041] Specifically, the charging state query unit 124 is configured to determine a status monitoring result of the battery module to be detected based on a feature interaction response pattern of the charging state query feature vector with respect to a set of battery module charging state timing correlation feature encoding vectors. Among them, Figure 4 is a block diagram of a charging state query unit in an energy management system according to an embodiment of the present application. As Figure 4 shown, the charging state query unit 124 includes: a charging state query response analysis subunit 1241, configured to input the charging state query feature vector and the set of battery module charging state timing correlation feature encoding vectors into a charging state single-point - global distribution state implicit modeling unit to obtain a charging state query response encoding vector of the battery module to be detected; a status monitoring result generation subunit 1242, configured to input the charging state query response encoding vector of the battery module to be detected into a charging state monitor based on a classifier to obtain the status monitoring result, where the status monitoring result is used to indicate whether there is an abnormality in the charging state.

[0042] More specifically, the charging state query response analysis subunit 1241 is configured to input the charging state query feature vector and the set of battery module charging state timing correlation feature encoding vectors into a charging state single-point - global distribution state implicit modeling unit to obtain a charging state query response encoding vector of the battery module to be detected. Here, in order to reveal whether there is an abnormal behavior deviating from the normal state during the charging process of the battery module to be detected, the present application further constructs a charging state single-point - global distribution state implicit modeling unit, and performs feature query response analysis on the charging state query feature vector of the battery module to be detected and the set of battery module charging state timing correlation feature encoding vectors, so as to capture potential difference features and overall distribution patterns between the charging state of the battery module to be detected and the features of other battery modules, and thus, based on the global charging state feature distribution of the battery module, more accurately identify the abnormal behavior of the battery module to be detected during the charging process.

[0043] Figure 5 is a block diagram of a charging state query response analysis subunit in an energy management system according to an embodiment of the present application. As Figure 5As shown, the charging status query response analysis subunit 1241 includes: a semantic response anchoring encoding secondary subunit 12411, configured to perform semantic response anchoring encoding on each charging status query feature vector in the set of the charging status query feature vector and the battery module charging status time-series association feature encoding vectors respectively to obtain a set of charging status feature semantic query response anchoring encoding matrices; a decision anchor adaptive splicing factor calculation secondary subunit 12412, configured to determine the decision anchor adaptive splicing factors of each charging status feature semantic query response anchoring encoding matrix based on the feature distribution of each charging status feature semantic query response anchoring encoding matrix in the set of the charging status feature semantic query response anchoring encoding matrices to obtain a set of charging status query response decision anchor adaptive splicing factors; and a feature fusion secondary subunit 12413, configured to fuse the set of the charging status feature semantic query response anchoring encoding matrices based on the set of the charging status query response decision anchor adaptive splicing weight factors to obtain the charging status query response encoding vector of the detected battery module.

[0044] In a specific example of the present application, the semantic response anchoring encoding secondary subunit 12411 is configured to: perform deep implicit feature extraction based on fully connected encoding on each battery module charging status time-series association feature encoding vector in the set of the charging status query feature vector and the battery module charging status time-series association feature encoding vectors respectively to obtain a set of charging status query feature deep implicit encoding vectors and battery module charging status time-series association feature deep implicit encoding vectors; and input each charging status query feature deep implicit encoding vector and each battery module charging status time-series association feature deep implicit encoding vector in the set of the battery module charging status time-series association feature deep implicit encoding vectors into a semantic response decision anchoring component respectively to obtain the set of the charging status feature semantic query response anchoring encoding matrices, which is expressed by the formula as:

[0045] V2 = {v 21 , v 22 ,..., v 2i ,..., v 2n}

[0046]

[0047] wherein, V1 represents the charging status query feature vector, V2 represents the set of the battery module charging status time-series association feature encoding vectors, v 21 , v 22 , v 2i and v 2nrespectively represent the 1st, 2nd, ith, and nth charging state time-series correlation feature coding vectors in the set of charging state time-series correlation feature coding vectors of the battery modules, where n is the number of vectors in the set of charging state time-series correlation feature coding vectors of the battery modules, represents matrix multiplication, W1 represents the charging state query feature weight matrix, W2 represents the charging state time-series correlation feature weight matrix of the battery modules, b1 represents the charging state query feature bias term, b2 represents the charging state time-series correlation feature bias term of the battery modules, sigoid(·) represents the sigmoid activation function, V d1 represents the charging state query feature deep implicit coding vector, v d2i represents v 2i corresponding charging state time-series correlation feature deep implicit coding vector of v, (·) T represents the transpose of a vector, S represents the feature scale factor, M 12i represents V d1 and v d2i charging state feature semantic query response anchoring coding matrix between them.

[0048] That is, in the present application, through the method of fully connected coding, the implicit semantic information of the charging state query feature vector and each charging state time-series correlation feature coding vector of the battery modules is deeply mined, so as to provide richer and more in-depth feature expressions for subsequent interaction analysis. Furthermore, the charging state query feature deep implicit coding vector and each charging state time-series correlation feature deep implicit coding vector of the battery modules are respectively subjected to semantic response decision anchoring coding, and the potential correlation features between the two are captured by calculating the outer product of vectors, so as to generate a set of charging state feature semantic query response anchoring coding matrices.

[0049] In a specific example of the present application, the decision anchor adaptive splicing factor calculation secondary subunit 12412 is used to: calculate the statistical eigenvalue of the charging state feature semantic query response anchoring coding matrix, and calculate its decision anchor adaptive splicing factor based on the statistical eigenvalue to obtain the charging state query response decision anchor adaptive splicing factor, where the statistical eigenvalue includes the maximum eigenvalue, eigenvalue variance, eigenvalue mean, and number of eigenvalues. More specifically, taking the sum of the eigenvalue variance and the drift coefficient of the charging state feature semantic query response anchoring coding matrix as the numerator, and calculating the square of the difference between the maximum eigenvalue of the charging state feature semantic query response anchoring coding matrix and its eigenvalue mean multiplied by its number of eigenvalues, plus the drift coefficient and twice the eigenvalue variance as the denominator, to obtain the charging state query response decision anchor adaptive splicing factor, which is expressed by the formula:

[0050] k = count(M 12i)

[0051]

[0052] Among them, count(·) represents calculating the number of elements of the matrix, k represents the difference amplification factor, that is, the number of eigenvalues of the charging state characteristic semantic query response anchored coding matrix, and σ 2 represents the characteristic variance of the charging state characteristic semantic query response anchored coding matrix, ∈ represents the drift coefficient of the charging state characteristic semantic query response anchored coding matrix, μ represents the characteristic mean of the charging state characteristic semantic query response anchored coding matrix, max(·) is the function of taking the maximum value, and E 12i represents the charging state query response decision anchor adaptive splicing factor corresponding to M 12i

[0053] That is, in order to more accurately measure the importance of semantic interaction information between different battery modules, the present application further introduces an adaptive weight allocation mechanism. By calculating the statistical eigenvalues of each charging state characteristic semantic query response anchored coding matrix, and based on these statistical eigenvalues, its decision anchor adaptive splicing factor is calculated to form a set of charging state query response decision anchor adaptive splicing factors, so as to measure the characteristic significance according to the characteristic distribution of each charging state characteristic semantic query response anchored coding matrix, and thus flexibly adjust the weights in the subsequent fusion process to improve the rationality and effectiveness of feature fusion.

[0054] In particular, here, the drift coefficient is used to smooth the fluctuation of the characteristic distribution of the charging state characteristic semantic query response anchored coding matrix. In a preferred example of the present application, for the drift coefficient ∈ of the charging state characteristic semantic query response anchored coding matrix, for the state transition of the eigenvalue set distribution of the charging state characteristic semantic query response anchored coding matrix from weak overall interpretability of the mean to strong local interpretability of the maximum value, the present application enhances the global dominance basis of the charging state characteristic semantic query response anchored coding matrix by introducing the weak-to-strong interpretable generalization of the drift coefficient ∈, which is expressed by the formula:

[0055]

[0056]

[0057] Among them, η is the intermediate transition representation value of the characteristic distribution balance state of the charging state characteristic semantic query response anchored coding matrix, and m ij is the j-th eigenvalue of matrix M 12i

[0058] ​​Here, η is used as an intermediate state transition representation from weak interpretability to strong interpretability. For each eigenvalue m of the charging state feature semantic query response anchored encoding matrix, ij it is used as the importance score of the charging state feature semantic query response anchored encoding matrix for global smooth state transition, to perform global control of the importance score weight of the intermediate state transition η relative to the global state transition, so as to realize the interpretable generalization inference of the weight basis of the charging state query response decision anchor adaptive splicing factor.

[0059] In a specific example of the present application, the feature fusion secondary subunit 12413 is used to: perform weight processing on the set of charging state query response decision anchor adaptive splicing factors based on the Softmax function to obtain a set of charging state query response decision anchor adaptive splicing weight factors; based on the set of charging state query response decision anchor adaptive splicing weight factors, perform weighted fusion and feature shape reshaping on the set of charging state feature semantic query response anchored encoding matrices to obtain the charging state query response encoding vector of the detected battery module, which is expressed by the formula:

[0060] a 12i = softmax(E 12i )

[0061]

[0062] where softmax(·) represents the normalized exponential function, a 12i represents the charging state query response decision anchor adaptive splicing weight factor of the matrix M 12i , M c represents the charging state feature semantic query response anchored encoding fusion matrix, reshape(·) represents the feature shape reshaping function, and v c represents the charging state query response encoding vector of the detected battery module.

[0063] That is, using the Softmax function to normalize the set of charging state query response decision anchor adaptive splicing factors into a weight set with probability distribution properties, and assigning a reasonable weight to each charging state feature semantic query response anchored encoding matrix to reflect its relative importance in the fusion process. Furthermore, based on the generated weight distribution, perform weighted fusion and feature shape reshaping on the set of charging state feature semantic query response anchored encoding matrices, so as to avoid the problems of information loss or feature confusion that may be brought by simple average fusion, and can perform targeted fusion according to the characteristics and importance of different charging state feature semantic query response anchored encoding matrices, so that the fused encoding vector can more comprehensively and accurately reflect the feature information of the charging state of the detected battery module.

[0064] More specifically, the state monitoring result generation subunit 1242 is configured to input the charging state query response encoding vector of the battery module to be detected into the classifier-based charging state monitor to obtain the state monitoring result, where the state monitoring result is used to indicate whether there is an abnormality in the charging state. Here, the charging state query response encoding vector of the battery module to be detected includes the charging state association pattern between the battery module to be detected and other battery modules during the charging process. After receiving the charging state query response encoding vector of the battery module to be detected, the classifier-based charging state monitor performs multi-layer feature learning on it to identify whether the charging state of the battery module to be detected deviates from the normal charging state feature distribution range, and then outputs the corresponding state monitoring result. If the state monitoring result shows that there is an abnormality in the charging state, the warning mechanism can be triggered in time to notify the management personnel to adjust the charging strategy to avoid performance degradation and potential safety risks caused by abnormal conditions such as overcharging, over-discharging or imbalance of the battery module.

[0065] In the above energy management system, the energy control module 130 is configured to determine the charging strategy for the battery module based on the situation of the power grid. It should be understood that the power grid, as the infrastructure for energy transmission and distribution, directly affects the charging process of the battery module. It should be understood that the load situation of the power grid is directly related to the stability and reliability of power supply. When the power grid load is too high, problems such as voltage drop and frequency fluctuation may occur, affecting the normal operation of the power grid. As an important load of the power grid, the charging strategy of the battery module needs to fully consider the real-time state of the power grid. Specifically, when the power grid is in a high-load state, in order to avoid causing additional pressure on the power grid, the charging power of the battery module can be appropriately reduced or the charging can be suspended; while when the power grid load is low and the power supply is sufficient, the charging power of the battery module can be increased to accelerate the charging speed. In this way, by real-time monitoring and analysis of the power grid state to dynamically adjust the charging strategy of the battery module, it helps to achieve the optimal utilization of energy, relieve the power supply pressure of the power grid, and then improve the stability and reliability of the power grid and promote the sustainable development of the energy system.

[0066] Specifically, when the power grid is in the low-load period, that is, when the electricity demand is relatively small, the charging rate of the battery module can be appropriately increased. At this time, since the power grid has sufficient capacity and the electricity price is usually low, increasing the charging intensity not only helps to balance the power grid load but also reduces the energy storage cost. In terms of specific operation, the energy control module sends instructions to the PCS (Power Conversion System) to adjust its working mode to achieve fast charging. At the same time, considering that fast charging may affect the battery life, it is also necessary to synchronously monitor parameters such as the temperature and voltage inside the battery to ensure that all indicators are within the safe range. Once an abnormal situation is detected, immediately reduce the charging power or suspend charging to prevent overheating or other safety hazards. On the contrary, if the power grid is in the peak-load stage, it means that the power supply is tense and the electricity cost rises. In this case, the charging activity of the battery module should be reduced or even stopped, and instead, it should be considered as a backup power source to participate in the power grid peak shaving. By releasing the stored energy to meet the emergency needs of some users, the power grid pressure can be relieved. To achieve this goal, the energy control module needs to accurately calculate the available electricity and allocate it to different load points according to the preset priority order. In addition, the possibility of working in coordination with other energy storage facilities can be explored to jointly build a flexible and variable energy scheduling network. For example, an energy storage device can be set near a solar power station. When the sun is shining brightly during the day, it gives priority to supplying power to local users and storing the excess electricity; after the sun sets in the evening, the energy storage device continues to provide power support to ensure the continuity of electricity use in the area is not affected.

[0067] In addition to coping with daily load fluctuations, a complete emergency plan is also needed for sudden power grid failures or extreme weather events. For example, when a strong storm hits and causes power outages in some areas, the energy storage system should quickly switch to the off-grid mode and independently take the responsibility of ensuring the normal operation of important facilities. In this special situation, the charging strategy naturally needs to be adjusted accordingly - suspend all non-essential charging tasks and concentrate resources to maintain the stable output of the existing reserve power. At the same time, closely track the changes in the external environment and resume the normal charging procedure immediately once the conditions permit. This can not only maximize the role of the energy storage device but also avoid causing new problems due to improper operation.

[0068] In summary, the energy management system based on the embodiments of the present application is described. It monitors the working status of battery modules by collecting the status parameters of each battery module in real time, and introduces a deep learning algorithm to perform time-series analysis of the charging status of each battery module in the charging state to capture the time-series charging characteristics of each battery module. Furthermore, based on the time-series charging characteristics of the global battery modules, it performs query analysis on the charging status of each battery module to achieve abnormal monitoring of the charging status of each battery module and charging strategy management. In this way, the charging status of each battery module can be obtained more effectively, potential charging anomalies can be discovered and processed in a timely manner, the charging process can be optimized, thereby improving the efficiency and reliability of the energy management system.

[0069] The basic principles of the present invention have been described above in conjunction with specific embodiments. However, it should be noted that the advantages, benefits, effects, etc. mentioned in the present invention are only examples and not limitations. It cannot be considered that these advantages, benefits, effects, etc. are essential for each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitating understanding, rather than limitations. The above details do not limit the present invention to necessarily adopt the above specific details for implementation.

[0070] In the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed or recorded in a certain embodiment, reference can be made to the relevant descriptions of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiments described above are only illustrative. For example, the unit division is only a logical function division, and there can be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0071] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present invention. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be encompassed by the present invention. Any associated drawing marks in the claims should not be regarded as limiting the claims involved.

[0072] In addition, it is obvious that the term "including" does not exclude other units or steps, and the singular does not exclude the plural. A plurality of units stated in the system claims can also be implemented by one unit through software or hardware.

[0073] Finally, it should be noted that the above description has been given for purposes of illustration and description. In addition, the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An energy management system, characterized in that: include: Data acquisition module, status monitoring module, energy control module; The data acquisition module is used to collect the state parameters of each battery module, and the state parameters include voltage, current and temperature; The state monitoring module is used to perform state monitoring on each battery module based on the state parameters of each battery module to obtain a set of state monitoring results, wherein the state monitoring on each battery module includes taking each battery module in a charging state as a monitoring object, and performing state query analysis on the battery module based on the charging characteristics of the global battery module to obtain the set of state monitoring results; The energy control module is used to determine a charging strategy for the battery module based on the status of the power grid.

2. The energy management system according to claim 1, characterized in that: The state monitoring module comprises: A charging state battery module information acquisition unit, used to extract the state parameters of the battery module in the charging state from the state parameters of the battery module to obtain a set of charging battery module state parameters, wherein the charging battery module state parameters include a time series of voltage, a time series of current and a time series of temperature; A charging state timing analysis unit, used for extracting charging state timing correlation features from each charging battery module state parameter in the set of charging battery module state parameters to obtain a set of charging state timing correlation feature coding vectors of the charging battery module; A query feature extraction unit, used to extract the charging state time series associated feature coding vector of the detected battery module from the set of charging state time series associated feature coding vectors of the charging state of the charging battery module as a charging state query feature vector; A charging state query unit is used to determine the state monitoring result of the detected battery module based on the characteristic interaction response mode of the charging state query feature vector relative to the set of battery module charging state time series associated feature coding vectors.

3. The energy management system according to claim 2, characterized in that: The charging state timing analysis unit is used for: A charging state timing encoder based on an LSTM model is used to perform timing correlation analysis on each charging battery module state parameter in the set of charging battery module state parameters to obtain a set of charging state timing correlation feature coding vectors of the charging battery module.

4. The energy management system according to claim 3, characterized in that: The charging status query unit comprises: A charging state query response analysis subunit, used for inputting the charging state query feature vector and the battery module charging state time series correlation feature coding vector set into the charging state single point-global distribution state implicit modeling unit to obtain the detected battery module charging state query response coding vector; The state monitoring result generating subunit is used to input the detected battery module charging state query response coding vector into the classifier-based charging state monitor to obtain the state monitoring result, and the state monitoring result is used to indicate whether there is an abnormality in the charging state.

5. The energy management system according to claim 4, characterized in that: The charging status query response analysis subunit includes: A semantic response anchor coding secondary subunit, used to perform semantic response anchor coding on each charging state query feature vector in the set of the charging state query feature vector and the battery module charging state time series associated feature coding vector to obtain a set of charging state feature semantic query response anchor coding matrices; A decision anchor adaptive splicing factor calculation secondary subunit is used to determine the decision anchor adaptive splicing factor of each charging state feature semantic query response anchor coding matrix in the set of charging state feature semantic query response anchor coding matrices based on the feature distribution of each charging state feature semantic query response anchor coding matrix to obtain a set of charging state query response decision anchor adaptive splicing factors; The feature fusion secondary subunit is used to fuse the set of charging state feature semantic query response anchor coding matrices based on the set of adaptive splicing weight factors of the charging state query response decision anchor to obtain the charging state query response coding vector of the detected battery module.

6. The energy management system according to claim 5, characterized in that: The semantic response anchor coding secondary subunit is used to: Performing deep implicit feature extraction based on fully connected coding on each battery module charging state time series associated feature coding vector in the set of the charging state query feature vector and the battery module charging state time series associated feature coding vector to obtain a set of charging state query feature deep implicit coding vectors and battery module charging state time series associated feature deep implicit coding vectors; Each battery module charging state time series associated feature deep implicit coding vector in the set of the charging state query feature deep implicit coding vector and the battery module charging state time series associated feature deep implicit coding vector is respectively input into the semantic response decision anchor component to obtain the set of the charging state feature semantic query response anchor coding matrix.

7. The energy management system according to claim 6, characterized in that: The decision anchor adaptive splicing factor calculation secondary subunit is used to: Calculate the statistical eigenvalues ​​of the charging status feature semantic query response anchor coding matrix, and calculate its decision anchor adaptive splicing factor based on the statistical eigenvalues ​​to obtain the charging status query response decision anchor adaptive splicing factor, wherein the statistical eigenvalues ​​include the maximum eigenvalue, the feature variance, the feature mean and the number of eigenvalues.

8. The energy management system according to claim 7, characterized in that: The decision anchor adaptive splicing factor calculation secondary subunit is used to: The sum of the feature variance and the drift coefficient of the charging state feature semantic query response anchor coding matrix is ​​used as the numerator, and the square of the difference between the maximum eigenvalue and the feature mean of the charging state feature semantic query response anchor coding matrix is ​​multiplied by the number of its eigenvalues, and then the drift coefficient and twice the feature variance are added as the denominator to obtain the charging state query response decision anchor adaptive splicing factor, wherein the drift coefficient is used to smooth the feature distribution fluctuation of the charging state feature semantic query response anchor coding matrix.

9. The energy management system according to claim 8, characterized in that: The feature fusion secondary subunit is used to: Performing a weighting process based on a Softmax function on the set of charging state query response decision anchor adaptive splicing factors to obtain a set of charging state query response decision anchor adaptive splicing weight factors; Based on the set of adaptive splicing weight factors of the charging state query response decision anchor, the set of charging state feature semantic query response anchor coding matrices is weighted fused and feature reshaped to obtain the charging state query response coding vector of the detected battery module.

Images