An energy storage power station EMS energy monitoring and management system

Abstract

The present invention relates to the field of energy monitoring and management technology, and specifically to an energy storage power station EMS energy monitoring and management system. It comprises: a data acquisition unit, a data processing unit, a data fusion unit, a battery status monitoring unit, an adaptive energy scheduling unit and an environmental monitoring unit. The data acquisition unit collects various data of the battery, and the data processing unit processes the data to construct a standard multi-dimensional operation data set. The data fusion unit extracts and fuses features of the data through a neural network to form a fused operation feature set. The battery status monitoring unit evaluates the battery based on its usage time and historical data, and outputs a battery status evaluation value. The adaptive energy scheduling unit optimizes the energy scheduling strategy based on the battery status and grid demand to ensure efficient energy distribution. In addition, the system also uses the environmental monitoring unit to collect environmental data in real time to further optimize the battery monitoring and energy scheduling strategy, and realize intelligent management of the energy storage power station.

Description

Technical Field

[0001] The present invention relates to the technical field of energy monitoring and management, and in particular to an energy storage power station EMS energy monitoring and management system. Background Art

[0002] With the continued growth of energy demand and the widespread use of renewable energy, power system stability, reliability, and energy efficiency face increasing challenges. To address these challenges, energy management systems (EMS) have emerged as a key technical means to optimize energy utilization, improve power system operational efficiency, and ensure power supply security. In power plants in particular, the energy monitoring and management functions of EMS are playing an increasingly important role in promoting the construction of smart grids and improving the economic efficiency and environmental sustainability of power plant operations.

[0003] The EMS system of a power plant is a core component of an energy storage plant, responsible for data collection, monitoring, analysis, and control of the battery management system, energy storage converter, and other related equipment. With the widespread adoption of renewable energy, the management and optimization of energy storage systems has become a key research area in the power industry. Specifically, the EMS system of an energy storage plant not only enables intelligent scheduling and optimized operation of energy storage facilities, but also improves the efficiency and reliability of the energy storage system, ensuring efficient energy use.

[0004] Existing EMS-based energy monitoring and management methods usually rely on a single data source (such as battery voltage, temperature, and capacity) for monitoring. However, in actual operation, the energy management of energy storage power stations requires a comprehensive analysis of multi-dimensional operating data such as battery performance, battery pack status, charge and discharge history, and ambient temperature and humidity to achieve accurate monitoring and management of the energy of energy storage power stations.

[0005] To this end, the present invention proposes an energy monitoring and management system for an energy storage power station (EMS). Summary of the Invention

[0006] The present invention aims to provide an energy storage power station EMS energy monitoring and management system, which includes: a data acquisition unit, a data processing unit, a data fusion unit, a battery status monitoring unit, an adaptive energy scheduling unit and an environmental monitoring unit; wherein the data acquisition unit provides basic data support for the system by collecting multi-dimensional operating data of the battery pack in real time; the data processing unit is responsible for processing the collected multi-dimensional operating data to construct a standardized data set, providing an accurate data source for subsequent analysis and decision-making; the data fusion unit combines multiple feature learning subnets with operating change characteristics to further optimize the battery pack status assessment and energy scheduling strategy; the battery status monitoring unit monitors the battery health status in real time by evaluating the deviation between the battery usage status and historical data; the adaptive energy scheduling unit intelligently adjusts the energy scheduling strategy of the energy storage system according to the battery status monitoring results and grid demand to achieve optimal energy utilization efficiency and cost-effectiveness; the environmental monitoring unit further optimizes the overall performance and operating efficiency of the energy storage system by conducting real-time monitoring of the surrounding environment and conducting comprehensive analysis with the multi-dimensional operating data of the battery pack. By integrating the functions of various modules, the present invention effectively improves the intelligent management level of the energy storage power station, ensures the stable operation of the battery pack, and optimizes energy scheduling to meet the dynamic needs of the power grid and reduce operating costs.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] An energy storage power station EMS energy monitoring and management system, comprising:

[0009] The data acquisition unit collects multi-dimensional operating data of each battery pack in the energy storage power station in real time, including: charging frequency, discharging frequency, voltage, current, temperature, humidity, load status and local grid voltage;

[0010] Furthermore, the multidimensional operation data is processed by a data processing unit, and a standard multidimensional operation data set is constructed;

[0011] Wherein, the data processing unit includes:

[0012] Cleaning the multidimensional operation data to obtain first multidimensional operation data;

[0013] Further, denoising is performed on the first multidimensional operation data to obtain second multidimensional operation data;

[0014] Further, filtering the second multi-dimensional operation data to obtain third multi-dimensional operation data;

[0015] Further, the third multi-dimensional operation data is formatted to obtain fourth multi-dimensional operation data;

[0016] Further, the fourth multi-dimensional operation data is standardized to obtain standard multi-dimensional operation data;

[0017] Further, a plurality of said standard multi-dimensional operation data are battery-pack labeled and integrated to obtain said standard multi-dimensional operation data set;

[0018] The standard multi-dimensional running data set is represented as: RMDS = {MD i |1≤i≤N}; where MD i is the standard multi-dimensional operating data of the i-th battery pack, and MD i =(v ij ,e ij ,t ij ,w ij ,s ij ,g i ), v ij is the standard voltage of the jth battery in the i-th battery pack; e ij is the standard current of the jth battery in the i-th battery pack; t ij is the standard battery temperature of the jth battery in the i-th battery pack; w ij is the standard battery humidity of the jth battery in the i-th battery pack; s ij is the standard battery load state of the jth battery in the i-th battery pack; g i is the standard grid local voltage of the ith battery pack; N is the total number of battery packs; M is the number of batteries.

[0019] Furthermore, the standard multi-dimensional operation data set is input into a data fusion unit for fusion to obtain a fused operation feature set;

[0020] Wherein, the data fusion unit includes:

[0021] Acquiring the standard multidimensional operating data set;

[0022] Further, the time interval for collecting each data in the standard multi-dimensional operation data set is calculated to obtain a time deviation vector;

[0023] Furthermore, the standard multi-dimensional operation data set and the time deviation vector are input into a trained neural network to obtain operation change characteristics;

[0024] Furthermore, a plurality of feature learning subnetworks are constructed according to the data types in the standard multidimensional operation data set; wherein the feature learning subnetwork includes: an input layer for receiving data in the standard multidimensional operation data set corresponding to the feature learning subnetwork; a feature extraction layer for extracting features from the data received by the input layer through one-dimensional convolution to obtain a feature vector; a feature change learning layer for combining the feature vector and the operation change feature learning change feature vector; wherein the formula of the feature change learning layer is: v t =σ(W*[f t :h t ]+b); where v t is the changing feature vector of time step t; σ() is the activation function; f t is the eigenvector of time step t; h t is the running change feature at time step t; [:] is the feature splicing operation; W is the weight matrix of the spliced ​​vector; b is the bias term;

[0025] Fusion weighting, for weighting the feature vector by the change feature vector to obtain a weighted feature vector;

[0026] An output layer, configured to output the weighted feature vector;

[0027] Furthermore, the weighted feature vector output by each of the feature learning subnetworks is input into a multi-scale fusion module to obtain the fusion running feature set.

[0028] Furthermore, the battery status monitoring unit is used to monitor the batteries in each battery pack;

[0029] Wherein, the battery status monitoring unit includes:

[0030] Get the expected usage time of each battery;

[0031] Furthermore, the current battery usage time is evaluated according to the expected usage time to obtain a battery usage status coefficient;

[0032] Furthermore, the deviation of various data between the current battery and the historical battery is calculated to obtain a deviation matrix;

[0033] Furthermore, the fusion operation feature set and the battery usage state coefficient are input into the state evaluation model, and the evaluation process of the state evaluation model is optimized using the deviation matrix to obtain a state evaluation value; wherein the calculation formula of the state evaluation value is: current =H(X,κ,P); where s currentis the current battery status evaluation value; H() is the function of the status evaluation model; X is the fusion operation feature set; κ is the battery usage status coefficient; P is the deviation matrix;

[0034] The state assessment model is trained using historical battery data and includes:

[0035] Obtaining historical battery data of the corresponding battery pack; wherein the historical battery data includes: historical battery usage status coefficient, historical multi-dimensional operation data, and historical battery status;

[0036] Further, the historical battery usage state coefficient and the historical multi-dimensional operation data are combined to obtain a training sample;

[0037] Furthermore, the training samples are received by the data input layer and processed to obtain standard training samples;

[0038] Furthermore, a convolutional layer is used to extract sample features of the standard training samples;

[0039] Furthermore, the sample features are input into the state association layer to obtain a state feature association matrix;

[0040] Furthermore, the sample features and the state feature association matrix are input into a feature association weighting layer to obtain weighted sample features;

[0041] Furthermore, the weighted sample features are input to the classification layer to obtain a battery status classification result;

[0042] Furthermore, the error between the battery state classification result and the historical battery state is calculated to optimize the state evaluation model; when the loss of the state evaluation model reaches convergence, the training ends.

[0043] Furthermore, according to the monitoring result of the battery status monitoring, energy scheduling is performed using an adaptive energy scheduling unit;

[0044] The adaptive energy scheduling unit includes:

[0045] According to the monitoring result of the battery status monitoring unit, a current state monitoring vector of the battery is constructed; wherein the current state monitoring vector is expressed as: (s current,ij |i=1,...,N;j=1,...,M); where s current,ij is the status evaluation value of the jth battery in the i-th battery pack; N is the total number of battery packs; M is the number of batteries;

[0046] Furthermore, the current state monitoring vector is used to calculate the current operating state evaluation value of the corresponding battery pack; wherein the calculation formula of the current operating state evaluation value is: Among them, R current,i is the current operating status evaluation value of the i-th battery pack; α j is the weight of the jth battery in the overall state assessment; β is the coefficient of the nonlinear interaction term between the control batteries; h(s current,ij ,s current,ik ) is a function of the nonlinear interaction between the j-th battery and the k-th battery;

[0047] Further, the current grid demand is obtained;

[0048] Further, a plurality of initial energy scheduling strategies are obtained according to the current operating state evaluation value of each battery pack and the current grid demand;

[0049] Furthermore, energy scheduling constraints for each battery pack are set; wherein the energy scheduling constraints are set according to the state evaluation value within the battery pack;

[0050] Further, the plurality of initial energy scheduling strategies are screened according to the energy scheduling constraint condition to obtain a set of candidate energy scheduling strategies;

[0051] Further, calculating the cost consumption and energy loss of the centralized implementation strategy of the candidate energy scheduling strategy;

[0052] Further, according to the comprehensive cost of the cost consumption and the energy loss;

[0053] Furthermore, the candidate energy scheduling strategy of the comprehensive cost is used as the optimal energy scheduling strategy.

[0054] In addition, the system also includes an environmental monitoring unit for real-time monitoring of environmental data around the energy storage, and comprehensive analysis of the environmental data with the multi-dimensional operating data of the battery pack to optimize the battery status monitoring unit and the adaptive energy scheduling unit.

[0055] Compared with the prior art, the present invention has the following beneficial effects:

[0056] 1. The present invention proposes a multi-dimensional data fusion method, which obtains a time deviation vector by obtaining a standard multi-dimensional operating data set and calculating the time interval for collecting each set of data. Then, the data set and the time deviation vector are input into the trained neural network to extract the operating change features. Next, a plurality of feature learning subnets are constructed according to the data type, and the convolution layer is used to extract the data features and the change features are learned in combination with the operating change features. Finally, the weighted feature vector of each data feature is obtained by weighted fusion, and the fused operating feature set is synthesized by a multi-scale fusion module. This method can effectively integrate various types of multi-dimensional and time-series data, thereby achieving more accurate feature recognition and prediction. For energy monitoring and management of energy storage power stations, the advantage of this method is that it can efficiently fuse multi-dimensional information from different data sources (such as battery status, load demand, and battery operating temperature, etc.) and monitor the operating status of the system in real time.

[0057] 2. The present invention proposes an energy monitoring method for monitoring the status of batteries, especially for dynamically evaluating and optimizing the operating status of battery packs in energy storage power stations. This method obtains the expected usage time of each battery and compares it with the current battery usage time, calculates the battery usage status coefficient, and then obtains the deviation matrix between the battery and historical data. Then, the current operating status of the battery is evaluated through a status evaluation model, and the model is optimized using a fusion of operating feature sets, status coefficients, and deviation matrices. This process can accurately evaluate the health status of each battery, provide real-time feedback on the battery status, and help predict the remaining life and efficiency attenuation of the battery. This method can significantly improve the energy monitoring accuracy of energy storage power stations and help optimize battery energy management and scheduling in long-term operations.

[0058] 3. This invention proposes an adaptive energy scheduling method for energy management in energy storage power plants (EMSs). This method monitors the health status of the battery pack in real time and constructs a current battery status monitoring vector to calculate the current operating status evaluation value of the battery pack. Based on the battery pack status evaluation value and the current grid demand, multiple initial energy scheduling strategies are generated. Energy scheduling constraints are set based on the battery pack status evaluation value to screen out a suitable set of candidate energy scheduling strategies. By comprehensively evaluating the cost consumption and energy loss of these strategies, the optimal energy scheduling strategy is selected to ensure efficient and stable energy management of the energy storage power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A structural diagram of the energy storage power station EMS energy monitoring and management system provided by an embodiment of the present invention;

[0060] Figure 2 A diagram illustrating the process of a data fusion unit provided in an embodiment of the present invention;

[0061] Figure 3 A comparison chart of monitoring experiments of the battery status monitoring unit provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0063] An EMS is a highly efficient system designed to improve energy efficiency, reduce costs, and ensure the reliability and sustainability of energy supply by integrating, controlling, and optimizing energy production, storage, distribution, and consumption. In energy storage power plants, the EMS is a core component responsible for data collection, monitoring, analysis, and control of the battery management system, energy storage converters, and other related equipment.

[0064] Existing EMS-based energy monitoring and management methods typically rely on a single data source (such as battery voltage, temperature, and capacity). However, in actual operation, energy management of energy storage power plants requires comprehensive consideration of multi-dimensional operational data, including battery performance, battery pack status, charge and discharge history, and ambient temperature and humidity. Only through comprehensive analysis of multi-dimensional data can accurate monitoring and optimized management of energy storage power plants be achieved.

[0065] To this end, the present invention proposes an advanced energy storage power station EMS energy monitoring and management system that can provide more accurate energy management strategies by integrating multi-dimensional data sources. The following will explain this in detail through two embodiments.

[0066] Example 1:

[0067] In the field of distributed energy, microgrids require efficient energy storage system management to optimize energy use and extend battery life. This embodiment designs an energy storage power station EMS energy monitoring and management system for microgrid scenarios. Figure 1 The system includes: a data acquisition unit, a data processing unit, a data fusion unit, a battery status monitoring unit, an adaptive energy scheduling unit and an environmental monitoring unit;

[0068] Among them, the data acquisition unit is used to collect multi-dimensional operating data of each battery pack in the energy storage power station in real time; the data processing unit is used to preprocess and standardize the collected multi-dimensional operating data; the data fusion unit is used to integrate the processed standard multi-dimensional operating data set and extract deep features; the battery status monitoring unit is used to monitor and evaluate the operating status of each battery pack in the energy storage power station; the adaptive energy scheduling unit is used to generate and execute the optimal energy scheduling strategy based on the battery status monitoring results and the current grid demand; the environmental monitoring unit is used to monitor the environmental data around the energy storage system in real time, and conduct comprehensive analysis in combination with the battery operating data to provide an environmental basis for system optimization.

[0069] Based on the functional descriptions of each unit above, the monitoring and management process for a microgrid energy storage power station is as follows: The system's data acquisition unit collects multi-dimensional operational data from each battery pack in the energy storage power station in real time. These batteries are widely used in distributed energy systems (such as wind and solar energy) to store excess renewable energy and provide power. This multi-dimensional data includes parameters such as charging frequency, discharging frequency, voltage, current, battery temperature, battery humidity, battery load status, and local grid voltage.

[0070] Furthermore, the data processing unit of the system is used to process the collected multidimensional data, and a standard multidimensional operation data set is constructed based on the processed data; wherein, Figure 1 The processing functions of the data processing unit include the following: cleaning the multi-dimensional operation data to obtain first multi-dimensional operation data;

[0071] Further, denoising is performed on the first multidimensional operation data to obtain second multidimensional operation data;

[0072] Further, filtering the second multi-dimensional operation data to obtain third multi-dimensional operation data;

[0073] Further, the third multi-dimensional operation data is formatted to obtain fourth multi-dimensional operation data;

[0074] Further, the fourth multi-dimensional operation data is standardized to obtain standard multi-dimensional operation data;

[0075] Furthermore, a plurality of standard multi-dimensional operation data are integrated according to the battery pack labels to obtain a standard multi-dimensional operation data set.

[0076] In the embodiments of the present application, the collected multi-dimensional operations are processed, including data cleaning, denoising, filtering, formatting, and standardization. This series of data processing processes can significantly improve the accuracy and reliability of energy storage power station energy monitoring and management. Cleaning and denoising ensure the validity and accuracy of the data; formatting and standardization ensure the uniformity and comparability of the data, facilitating subsequent analysis and decision-making. Ultimately, the integrated standard multi-dimensional operation data set can provide a more accurate and real-time monitoring and analysis basis for the energy management system, thereby optimizing the operating efficiency of the energy storage power station.

[0077] For a standard multidimensional running data set, it is expressed as: RMDS = {MD i |1≤i≤N}; where MD i is the standard multi-dimensional operating data of the i-th battery pack, and MD i =(v ij ,e ij ,t ij ,w ij ,s ij ,g i ), v ij is the standard voltage of the jth battery in the i-th battery pack; e ij is the standard current of the jth battery in the i-th battery pack; t ij is the standard battery temperature of the jth battery in the i-th battery pack; w ij is the standard battery humidity of the jth battery in the i-th battery pack; s ij is the standard battery load state of the jth battery in the i-th battery pack; g i is the standard grid local voltage of the ith battery pack; N is the total number of battery packs; M is the number of batteries.

[0078] The data content of a standardized multidimensional operational dataset is described in the examples of this application. This data structure is of great significance for energy monitoring and management in energy storage power plants. This standardized multidimensional operational dataset provides accurate real-time monitoring information for each battery pack and its individual cells. This enables the energy management system to implement precise control and optimization based on key information such as voltage, current, temperature, and humidity for each cell.

[0079] The energy storage power station of the embodiment of the present application has a total of 4 battery packs, and each battery pack contains at least 5 batteries with different storage flux and power; therefore, the value of N in the standard multi-dimensional operation data set is 4, and the value of M is determined according to the number of batteries in the battery pack. The standard multi-dimensional operation data set is input into the data fusion unit for fusion to obtain a fusion operation feature set; refer to Figure 2 , Figure 2The fusion process of the data fusion unit is described as follows: obtaining a standard multi-dimensional operating data set; referring to Table 1, which shows the standard operating data of battery No. 1 and battery No. 2 in battery pack No. 1;

[0080] Table 1 Standard operating data of No. 1 battery and No. 2 battery in No. 1 battery pack

[0081]

[0082] Furthermore, the interval time of the same parameter change of each data of the same battery in the instantaneous interval in the standard multi-dimensional operation data set is calculated; the corresponding number of time intervals can be obtained according to the data collection amount, and the time deviation vector is obtained;

[0083] Furthermore, the time deviation vector and the standard multidimensional dataset are input into the trained neural network to obtain time-based operational variation characteristics; wherein the operational variation characteristics refer to the charge variation, discharge variation, voltage variation, current variation, battery temperature variation, battery humidity variation, battery load state variation of each battery, and the local voltage variation of the power grid of the battery pack; in addition, the neural network proposed in this application example refers to a long short-term memory neural network model;

[0084] Furthermore, multiple feature learning subnetworks are constructed based on the data types in the standard multidimensional running dataset; Figure 2 It can be seen from the content that the feature learning subnets in the embodiment of the present application include: charging frequency feature learning subnet, discharging frequency feature learning subnet, voltage feature learning subnet, current feature learning subnet, battery temperature feature learning subnet, battery humidity feature learning subnet, battery load status feature learning subnet and power grid local voltage feature learning subnet.

[0085] Furthermore, the standard multi-dimensional running data set is input into the corresponding feature learning subnetwork to obtain multiple weighted feature vectors;

[0086] Among them, the feature learning subnet includes:

[0087] The input layer is used to receive data from the standard multi-dimensional running dataset corresponding to the feature learning subnetwork;

[0088] The feature extraction layer is used to extract features from the data received from the input layer through one-dimensional convolution to obtain a feature vector;

[0089] The feature change learning layer is used to combine the feature vector and the running change feature to learn the change feature vector; the formula of the feature change learning layer is: t =σ(W*[f t :h t ]+b); where v tis the changing feature vector of time step t; σ() is the activation function; f t is the eigenvector of time step t; h t is the running change feature at time step t; [:] is the feature splicing operation; W is the weight matrix of the spliced ​​vector; b is the bias term;

[0090] The fusion weighting layer is used to weight the feature vector by the changing feature vector to obtain the weighted feature vector. The fusion weighting layer uses the attention mechanism to focus on the features with larger changing trends in the changing feature vector. The weighting formula of the weighted feature vector is: weighted =α t ⊙f t ;f weighted is the weighted eigenvector; ⊙ is the dot multiplication formula; α t is the feature weight of time step t, that is Where score( ) is the function of evaluating feature importance using additive attention; T is the time step of the entire data collection; exp( ) is the exponential function;

[0091] Output layer, outputs weighted feature vector;

[0092] Furthermore, the weighted feature vectors output by multiple feature learning subnetworks are input into the multi-scale fusion module to obtain the fused running feature set.

[0093] Among them, in the multi-scale fusion module, each weighted feature vector is upgraded to the same dimension and then fused into a fusion operation feature set through splicing.

[0094] Further, using Figure 1 The battery status monitoring unit in the system monitors each battery;

[0095] Among them, the battery status monitoring unit specifically

[0096] Obtaining the expected service life of each battery in the battery pack; wherein the expected service life is calculated by the standard service life of the battery and the attenuation of the battery life during use;

[0097] Furthermore, the current battery usage time is evaluated according to the expected usage time to obtain a battery usage status coefficient. The battery usage status coefficient is calculated as follows: Among them, C status T is the battery usage status coefficient; actual The current battery life; T expected The expected duration of use;

[0098] Furthermore, the deviations of various data between the current battery and the historical battery are calculated to obtain a deviation matrix; wherein the deviation matrix includes: battery performance deviation, battery capacity deviation, battery cycle life deviation, and battery internal resistance deviation;

[0099] Furthermore, the fused operation feature set and the battery usage status coefficient are input into the state assessment model;

[0100] The state assessment model is trained using historical battery data. The specific training process includes:

[0101] Obtain historical battery data of the corresponding battery pack, including: battery usage status coefficient, historical multi-dimensional operation data and historical battery status;

[0102] Further, the historical battery usage state coefficient and the historical multi-dimensional operation data are combined to obtain a training sample;

[0103] Furthermore, the training samples are received by the data input layer and processed to obtain standard training samples;

[0104] Furthermore, the convolutional layer is used to extract sample features of the standard training samples;

[0105] Furthermore, the sample features are input into the state association layer to obtain the state feature association matrix. The state association layer is used to calculate the correlation between the interaction of each feature in the sample and the battery state prediction. The specific calculation process is as follows:

[0106] (1) Calculate the local correlation matrix, the calculation formula is: R local (F i ,F j )=α local,i,j *g(F i ,F j ); where R local (F i ,F j ) is the sample feature F i With sample feature F j The local correlation degree of F i and F j are sample features of the same feature type; α local,i,j is the attention weight, which indicates the importance between local features; g() is the feature interaction function;

[0107] (2) Calculate the global correlation matrix, the calculation formula is: R global (F i ,F k )=α global,i,k *g(F i ,F k ); where Rglobal (F i ,F k ) is the sample feature F i With sample feature F k The global correlation of F i and F k are the characteristics of samples of different feature types; α global,i,k is the attention weight, which indicates the importance between global features;

[0108] (3) The local correlation matrix and the global correlation matrix are fused, and the fused state correlation matrix is ​​expressed as: Among them, R local The local correlation matrix obtained by integrating all local correlation degrees; R global The global correlation matrix obtained by integrating all global correlations; is the balance coefficient, which is used to adjust the importance of the local and global relationship;

[0109] As shown in Table 2, the state feature correlation matrix in the embodiment of the present application is given in Table 2; the details are as follows:

[0110] Table 2 State feature correlation matrix

[0111]

[0112] Furthermore, the sample features and state feature association matrix are input into the feature association weighting layer to obtain weighted sample features;

[0113] Furthermore, the weighted sample features are input to the classification layer to obtain the battery status classification results; wherein the battery status classification results are divided into: normal state, slight abnormality, serious abnormality and fault state;

[0114] Furthermore, the error between the battery status classification result and the historical battery status is calculated to optimize the status evaluation model; when the loss of the status evaluation model reaches convergence, the training ends.

[0115] In the embodiments of this application, a condition assessment model is trained using historical battery data. This method enables the trained condition assessment model to accurately predict battery health and performance changes during actual battery use, thereby providing more scientific battery management decisions. This benefit is that the model can comprehensively learn from historical data, thereby improving the accuracy of battery condition predictions, extending battery life, optimizing battery maintenance strategies, and effectively avoiding premature or late maintenance interventions, ensuring efficient and safe battery operation.

[0116] Furthermore, the deviation matrix is ​​used to optimize the evaluation process of the state evaluation model to obtain the state evaluation value; wherein, the calculation formula of the state evaluation value is: scurrent =H(X,κ,P); where s current is the current battery status evaluation value; H() is the function of the status evaluation model; X is the fusion operation feature set; κ is the battery usage status coefficient; P is the deviation matrix;

[0117] In order to illustrate the monitoring effect of the battery status monitoring unit in the embodiment of the present application, an experiment was carried out using historical battery data. In this experiment, the state evaluation model and the state evaluation model after the deviation matrix optimization were compared with the actual battery state; see Table 3, which shows the results of 10 comparisons, and Figure 3 The evaluation accuracy during training is given in .

[0118] Table 3. Result description of 10 battery status comparisons

[0119] Experiment number State assessment model evaluation value Evaluation value of the optimized state evaluation model Real battery status value 1 0.85 0.86 0.87 2 0.60 0.63 0.62 3 0.45 0.49 0.50 4 0.72 0.74 0.73 5 0.90 0.92 0.91 6 0.55 0.57 0.56 7 0.50 0.52 0.51 8 0.78 0.79 0.79 9 0.65 0.66 0.67 10 0.40 0.41 0.41

[0120] Furthermore, the adaptive energy scheduling unit is used to perform energy scheduling according to the current battery status. The specific process includes:

[0121] According to the monitoring result of the battery status monitoring unit, the current status monitoring vector of the battery is constructed; wherein the current status monitoring vector is expressed as: (s current,ij |i=1,...,N;j=1,...,M); where s current,ij is the status evaluation value of the jth battery in the i-th battery pack; N is the total number of battery packs; M is the number of batteries;

[0122] Furthermore, the current state monitoring vector is used to calculate the current operating state evaluation value of the corresponding battery pack; wherein the calculation formula of the current operating state evaluation value is: Among them, R current,i is the current operating status evaluation value of the i-th battery pack; α j is the weight of the jth battery in the overall state assessment; β is the coefficient of the nonlinear interaction term between the control batteries; h(s current,ij ,s current,ik ) is a function of the nonlinear interaction between the j-th battery and the k-th battery;

[0123] Further, the current grid demand is obtained;

[0124] Further, a plurality of initial energy scheduling strategies are obtained according to the current operating state evaluation value of each battery pack and the current grid demand;

[0125] Furthermore, energy scheduling constraints for each battery pack are set; wherein the energy scheduling constraints are set according to the state evaluation value within the battery pack;

[0126] Further, the plurality of initial energy scheduling strategies are screened according to the energy scheduling constraint condition to obtain a set of candidate energy scheduling strategies;

[0127] Further, calculating the cost consumption and energy loss of the centralized implementation strategy of the candidate energy scheduling strategy;

[0128] Further, according to the comprehensive cost of the cost consumption and the energy loss;

[0129] Furthermore, the candidate energy scheduling strategy of the comprehensive cost is used as the optimal energy scheduling strategy.

[0130] In this embodiment of the application, an adaptive energy scheduling unit dynamically adjusts the battery pack's operating status and energy scheduling strategy through real-time battery status monitoring and analysis, thereby accurately matching grid demand, optimizing energy distribution, and maximizing battery efficiency and grid operational stability. This adaptive energy scheduling method intelligently optimizes the energy scheduling strategy based on the battery pack's real-time health status and grid demand, thereby improving grid efficiency and reliability.

[0131] In addition, Figure 1 The system includes an environmental monitoring unit for real-time monitoring of environmental data around the energy storage system and comprehensive analysis of the environmental data with the multi-dimensional operating data of the battery pack to optimize the battery status monitoring unit and the adaptive energy scheduling unit;

[0132] The optimization of the battery status monitoring unit and the adaptive energy scheduling unit for environmental data includes:

[0133] (1) Battery Status Monitoring Unit Optimization: By comprehensively analyzing environmental data (such as temperature, humidity, air pressure, and light intensity) with battery pack operating data (such as battery voltage, current, temperature, and charge and discharge cycles), the potential impact of environmental changes on battery health can be identified. For example, excessively high or low temperatures can directly affect the battery's charge and discharge efficiency and lifespan. The environmental monitoring unit can capture this data in real time and feed it back to the battery status monitoring unit, thereby adjusting the battery status assessment model to ensure more accurate battery status assessment.

[0134] (2) Adaptive Energy Scheduling Unit Optimization: Real-time environmental data provided by the environmental monitoring unit can help the adaptive energy scheduling unit perform more accurate energy scheduling. For example, when the external ambient temperature is low, the battery charging efficiency may decrease. The system can adjust the scheduling strategy to avoid overcharging or discharging, ensuring that the battery operates within the optimal operating temperature range. By combining environmental data with battery operating data, the battery's charge and discharge cycle can be adjusted more intelligently to reduce energy loss and extend the battery's service life.

[0135] In an embodiment of the present application, the environmental monitoring unit collects real-time environmental data surrounding the energy storage system and analyzes it in combination with the battery pack's operating data, thereby achieving accurate assessment of battery status and intelligent optimization of energy scheduling. Specifically, the environmental monitoring unit combines environmental factors such as temperature, humidity, air pressure, and light intensity with operating data such as battery voltage, current, temperature, and charge and discharge cycles to provide real-time feedback to the battery status monitoring unit to adapt to battery health assessments under different environmental conditions. At the same time, real-time environmental data can also help the adaptive energy scheduling unit accurately adjust the charge and discharge strategy to avoid efficiency loss or performance degradation of the battery due to environmental factors.

[0136] In an embodiment of the present application, the energy monitoring and management system of the energy storage power station EMS (energy management system) proposed in the present invention is adopted to realize comprehensive monitoring and efficient management of the energy storage power station in the microgrid. The system collects the multi-dimensional operating data of each battery group in the energy storage power station in real time through the data acquisition unit, and processes these data using the data processing unit to construct a standard multi-dimensional operating data set, and then fuses the data through the data fusion unit to obtain a fused operating feature set. The battery status monitoring unit accurately monitors the battery based on the fused operating feature set and the battery usage status coefficient, and calculates the current battery status evaluation value by optimizing the status evaluation model. Finally, the system uses the adaptive energy scheduling unit to perform energy scheduling according to the monitoring results to ensure that the energy storage power station operates in the optimal state.

[0137] Example 2:

[0138] In Example 1, the energy storage station EMS energy monitoring and management system of the present invention is used to achieve efficient monitoring and intelligent management of microgrid energy storage systems in the field of distributed energy. To further illustrate the effectiveness of the present invention, the energy storage station EMS energy monitoring and management system is applied to a photovoltaic energy storage system in this embodiment of the application. The specific process includes:

[0139] Use the data acquisition unit to collect multi-dimensional operating data of each battery pack in the photovoltaic energy storage power station in real time;

[0140] Furthermore, the multidimensional operation data is processed by a data processing unit, and a standard multidimensional operation data set is constructed;

[0141] Furthermore, the standard multi-dimensional operation data set is input into the data fusion unit for fusion to obtain a fused operation feature set;

[0142] Furthermore, the weighted feature vector output by each feature learning subnetwork is input into the multi-scale fusion module to obtain the fusion running feature set.

[0143] Furthermore, the battery status monitoring unit is used to monitor the batteries in each battery pack;

[0144] The battery status monitoring unit includes:

[0145] Get the expected usage time of each battery;

[0146] Furthermore, the current battery usage time is evaluated according to the expected usage time to obtain a battery usage status coefficient;

[0147] Furthermore, the deviation of various data between the current battery and the historical battery is calculated to obtain a deviation matrix;

[0148] Furthermore, the fusion operation feature set and the battery usage state coefficient are input into the state evaluation model, and the evaluation process of the state evaluation model is optimized using the deviation matrix to obtain the state evaluation value; wherein, the calculation formula of the state evaluation value is: s current =H(X,κ,P); where s current is the current battery status evaluation value; H() is the function of the status evaluation model; X is the fusion operation feature set; κ is the battery usage status coefficient; P is the deviation matrix;

[0149] Furthermore, according to the monitoring result of the battery status monitoring, energy scheduling is performed using an adaptive energy scheduling unit;

[0150] The adaptive energy scheduling unit includes:

[0151] According to the monitoring results of the battery status monitoring unit, the current state monitoring vector of the battery is constructed; wherein the current state monitoring vector is expressed as: (s current,ij |i=1,...,N;j=1,...,M); where s current,ij is the status evaluation value of the jth battery in the i-th battery pack; N is the total number of battery packs; M is the number of batteries;

[0152] Furthermore, the current state monitoring vector is used to calculate the current operating state evaluation value of the corresponding battery pack; wherein the calculation formula of the current operating state evaluation value is: Among them, R current,i is the current operating status evaluation value of the i-th battery pack; α j is the weight of the jth battery in the overall state assessment; β is the coefficient of the nonlinear interaction term between the control batteries; h(s current,ij ,s current,ik ) is a function of the nonlinear interaction between the j-th battery and the k-th battery;

[0153] Further, the current grid demand is obtained;

[0154] Furthermore, multiple initial energy scheduling strategies are obtained based on the current operating status evaluation value of each battery group and the current grid demand;

[0155] Furthermore, energy scheduling constraints are set for each battery pack; wherein the energy scheduling constraints are set according to the state evaluation value within the battery pack;

[0156] Furthermore, multiple initial energy scheduling strategies are screened according to energy scheduling constraints to obtain a set of candidate energy scheduling strategies;

[0157] Further, calculating the cost consumption and energy loss of the centralized implementation strategy of the candidate energy scheduling strategy;

[0158] Further, according to the comprehensive cost of cost consumption and said energy loss;

[0159] Furthermore, the candidate energy scheduling strategy with comprehensive cost is taken as the optimal energy scheduling strategy.

[0160] Furthermore, the environmental monitoring unit is used to monitor the environmental data around the energy storage in real time, and the environmental data is comprehensively analyzed with the multi-dimensional operating data of the battery pack to optimize the battery status monitoring unit and the adaptive energy scheduling unit.

[0161] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. An energy storage power station EMS energy monitoring and management system, characterized in that: The system comprises: The data acquisition unit collects multi-dimensional operating data of each battery pack in the energy storage power station in real time, including: charging frequency, discharging frequency, voltage, current, battery temperature, battery humidity, battery load status and local grid voltage; Processing the multidimensional operation data by a data processing unit and constructing a standard multidimensional operation data set; Inputting the standard multi-dimensional operation data set into a data fusion unit for fusion to obtain a fusion operation feature set; Using a battery status monitoring unit to monitor the batteries in each battery pack; Wherein, the battery status monitoring unit includes: Get the expected usage time of each battery; Evaluate the current battery usage time according to the expected usage time to obtain a battery usage status coefficient; Calculate the deviation of various data between the current battery and the historical battery to obtain the deviation matrix; The fusion operation feature set and the battery usage state coefficient are input into the state evaluation model, and the evaluation process of the state evaluation model is optimized using the deviation matrix to obtain a state evaluation value; wherein the calculation formula of the state evaluation value is: current =H(X,κ,P); where s current is the current battery status evaluation value; H() is the function of the status evaluation model; X is the fusion operation feature set; κ is the battery usage status coefficient; P is the deviation matrix; According to the monitoring result of the battery status monitoring unit, energy scheduling is performed using the adaptive energy scheduling unit.

2. The energy storage power station EMS energy monitoring and management system according to claim 1 is characterized in that: The data processing unit includes: cleaning the multidimensional operation data to obtain first multidimensional operation data; denoising the first multidimensional operation data to obtain second multidimensional operation data; filtering the second multidimensional operation data to obtain third multidimensional operation data; formatting the third multidimensional operation data to obtain fourth multidimensional operation data; standardizing the fourth multidimensional operation data to obtain standard multidimensional operation data; battery pack labeling and integration of multiple standard multidimensional operation data to obtain the standard multidimensional operation data set.

3. The energy storage power station EMS energy monitoring and management system according to claim 1 is characterized in that: The standard multi-dimensional running data set is expressed as: RMDS = {MD i |1≤i≤N}; Among them, MD i is the standard multi-dimensional operating data of the i-th battery pack, and MD i =(v ij ,e ij ,t ij ,w ij ,s ij ,g i ), v ij is the standard voltage of the jth battery in the i-th battery pack; e ij is the standard current of the jth battery in the i-th battery pack; t ij is the standard battery temperature of the jth battery in the i-th battery pack; w ij is the standard battery humidity of the jth battery in the i-th battery pack; s ij is the standard battery load state of the jth battery in the i-th battery pack; g i is the standard grid local voltage of the ith battery pack; N is the total number of battery packs; M is the number of batteries.

4. The energy storage power station EMS energy monitoring and management system according to claim 1, characterized in that: The data fusion unit includes: Acquiring the standard multidimensional operating data set; Calculating the time interval for collecting each piece of data in the standard multi-dimensional operation data set to obtain a time deviation vector; Inputting the standard multi-dimensional operation data set and the time deviation vector into a trained neural network to obtain operation change characteristics; Constructing multiple feature learning subnetworks according to the data types in the standard multidimensional operating data set; wherein the feature learning subnetworks include: An input layer, configured to receive data in the standard multidimensional operating dataset corresponding to the feature learning subnet; A feature extraction layer, configured to extract features from the data received from the input layer through one-dimensional convolution to obtain a feature vector; A feature change learning layer is used to combine the feature vector and the operation change feature to learn the change feature vector; wherein the formula of the feature change learning layer is: t =σ(W*[f t :h t ]+b); where v t is the changing feature vector of time step t; σ() is the activation function; f t is the feature vector of time step t; h t is the running change feature at time step t; [:] is the feature splicing operation; W is the weight matrix of the spliced ​​vector; b is the bias term; Fusion weighting, for weighting the feature vector by the change feature vector to obtain a weighted feature vector; An output layer, configured to output the weighted feature vector; The weighted feature vector output by each feature learning subnet is input into a multi-scale fusion module to obtain the fusion operation feature set.

5. The energy storage power station EMS energy monitoring and management system according to claim 1, characterized in that: The state assessment model is trained using historical battery data and includes: Obtaining historical battery data of the corresponding battery pack; wherein the historical battery data includes: historical battery usage status coefficient, historical multi-dimensional operation data, and historical battery status; Combining the historical battery usage state coefficient and the historical multi-dimensional operation data to obtain a training sample; The training samples are received by the data input layer and processed to obtain standard training samples; Extracting sample features of the standard training samples using a convolutional layer; Inputting the sample features into the state association layer to obtain a state feature association matrix; Inputting the sample features and the state feature association matrix into a feature association weighting layer to obtain weighted sample features; Inputting the weighted sample features into the classification layer to obtain a battery status classification result; The error between the battery state classification result and the historical battery state is calculated to optimize the state evaluation model; when the loss of the state evaluation model reaches convergence, the training ends.

6. The energy storage power station EMS energy monitoring and management system according to claim 1, characterized in that: The adaptive energy scheduling unit includes: According to the monitoring result of the battery status monitoring unit, a current state monitoring vector of the battery is constructed; wherein the current state monitoring vector is expressed as: (s current,ij |i=1,...,N;j=1,...,M); where s current,ij is the status evaluation value of the jth battery in the i-th battery pack; N is the total number of battery packs; M is the number of batteries; The current state monitoring vector is used to calculate the current operating state evaluation value of the corresponding battery pack; wherein the calculation formula of the current operating state evaluation value is: Among them, R current,i is the current operating status evaluation value of the i-th battery pack; α j is the weight of the jth battery in the overall state assessment; β is the coefficient of the nonlinear interaction term between the control batteries; h(s current,ij ,s current,ik ) is a function of the nonlinear interaction between the j-th battery and the k-th battery; Obtain current grid demand; Obtaining a plurality of initial energy scheduling strategies according to the current operating state evaluation value of each battery group and the current grid demand; Setting energy scheduling constraints for each battery pack; wherein the energy scheduling constraints are set according to the state evaluation value within the battery pack; Screening the plurality of initial energy scheduling strategies according to the energy scheduling constraint condition to obtain a set of candidate energy scheduling strategies; Calculate the cost consumption and energy loss of the centralized implementation strategy of the candidate energy scheduling strategy; Comprehensive cost based on the cost consumption and the energy loss; The candidate energy scheduling strategy corresponding to the minimum comprehensive cost is taken as the optimal energy scheduling strategy.

7. The energy storage power station EMS energy monitoring and management system according to claim 1, characterized in that: The system also includes an environmental monitoring unit for monitoring the environmental data around the energy storage in real time, and performing a comprehensive analysis of the environmental data with the multi-dimensional operating data of the battery pack to optimize the battery status monitoring unit and the adaptive energy scheduling unit.

Images