Battery energy storage power station status assessment and intelligent perception method and system

Abstract

The state assessment and intelligent perception method and system of a battery energy storage power station first collect the core monitoring quantities of the battery energy storage power station, and perform data cleaning and preprocessing. Then, through multi-order feature extraction and complex relationship fusion, the multi-dimensional monitoring quantity information is mapped into the state assessment space to achieve comprehensive quantification of the battery operating status. Next, a deep time feature convolutional network is used to model the evolution trajectory of the battery pack state points in the state assessment space. Finally, through the fuzzy expert state analysis system, the health status and abnormal event information of the battery energy storage power station are output to provide a basis for the situational awareness of the battery energy storage power station. The present invention can solve the problem of insufficient operational situation awareness, health status assessment and operational trend prediction capabilities of battery energy storage power stations, and provide technical support for the intelligent operation and maintenance of battery energy storage power stations.

Description

Technical Field

[0001] The present invention relates to the technical field of battery status assessment, and in particular to a method and system for battery energy storage power station status assessment and intelligent perception. Background Art

[0002] With the rapid development of new energy technologies, battery energy storage technology has gradually become an important means of addressing power system stability issues. Battery energy storage power stations achieve dynamic grid load balance by regulating the battery's charge and discharge status, thereby ensuring stable grid operation. However, over the long term, battery energy storage power stations may experience varying degrees of performance degradation or even failure due to factors such as environmental impact, battery aging, and complex operating conditions. Therefore, it is necessary to comprehensively and in real time perceive the operational status of battery energy storage power stations, as well as to quantitatively assess their health and predict trends.

[0003] In the existing technology, most of the evaluation methods for the health status of battery energy storage power stations are based on the analysis of a single parameter or traditional machine learning algorithms. However, due to the complexity of battery energy storage power stations, these methods can usually only evaluate the status of a single battery cluster. It is difficult to integrate multi-dimensional monitoring data to capture the complex coupling relationship of the overall operation of the energy storage power station, and it is unable to accurately reflect the overall operating status of the energy storage power station. The patent with publication number CN119024195A proposes an online evaluation method for digital energy storage battery packs. It uses a digital data acquisition and transmission system to collect the operating parameter data of the energy storage battery pack in real time, and uses a deep learning algorithm to build a data analysis model system. It automatically identifies the abnormal status of the energy storage battery pack, predicts the changing trend of its future performance indicators, and generates an evaluation report, thereby realizing intelligent monitoring and management of the energy storage battery pack. The patent with publication number CN114578251A proposes a battery module safety status assessment method and device based on a convolutional neural network. The method includes extracting the output voltage, current and surface temperature of the battery cell under different states, forming a database and dividing it into a training set and a test set, using it to train the convolutional neural network until the output accuracy reaches the standard, and then inputting real-time data to obtain the safety and health status information of the battery cell. The problems with these methods are: (1) the data dimension is low, and key aging indicators such as internal resistance and capacity are not integrated, resulting in a one-sided state assessment; (2) the complex coupling relationship between multiple parameters is not explicitly modeled, making it difficult to reflect the overall operating status of the battery pack; (3) the battery packs of the energy storage power station are not comprehensively analyzed, and the long-term dynamic evolution law of the energy storage power station cannot be predicted. In summary, the existing methods only perform local assessments on the battery packs, and cannot achieve quantitative assessment and intelligent perception of the overall operating status of the battery energy storage power station, and have the core defect of "seeing the trees but not the forest".

[0004] Due to the lack of intelligent health status assessment and trend prediction methods, energy storage power stations still face challenges in terms of safety, reliability, and operational efficiency. This restricts the promotion and application of the "condition-based maintenance" model and hinders the large-scale deployment of energy storage power stations in new power systems. Therefore, there is an urgent need for an intelligent perception and status assessment method for the operating status of battery energy storage power stations based on multi-parameter fusion analysis that can comprehensively capture complex coupling relationships. This method can improve the intelligent operation and maintenance level and operational safety of energy storage power stations and promote the efficient development of new power systems. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method and system for battery energy storage power station status assessment and intelligent perception based on multi-parameter fusion analysis, capable of comprehensively capturing complex coupling relationships. This invention addresses the inadequate operational status awareness, health status assessment, and operational trend prediction capabilities of battery energy storage power stations, providing technical support for the intelligent operation and maintenance of battery energy storage power stations.

[0006] The present invention adopts the following technical solutions.

[0007] The present invention proposes a battery energy storage power station status assessment and intelligent perception method, including:

[0008] Step S10: Collect historical data of various monitoring quantities in the battery energy storage power station, including the voltage of each battery pack , current ,capacity , internal resistance and temperature , and perform data cleaning and preprocessing;

[0009] Step S20: Input the preprocessed data into a time-frequency domain hybrid feature extraction module to construct a multi-order feature vector of the monitoring quantity; extract the complex coupling relationship between the monitoring quantities in the multi-order feature vector through a relationship fusion module; generate a state assessment vector based on the complex coupling relationship to quantify the operating status of the battery pack of the battery energy storage power station;

[0010] Step S30: construct a state evaluation space based on the state evaluation vector ; In the state evaluation space In the process, the state point evolution trajectory of the battery pack of the battery energy storage power station is modeled by a deep temporal feature convolutional network to obtain the high-order state vector in the state point evolution trajectory. The modeling formula is as follows:

[0011] ;

[0012] in, Indicates the The high-order state vector obtained by time step prediction; Represents the deep temporal feature convolutional network DTFCN; Indicates the The state evaluation vector corresponding to the time step; Represents the local trajectory adaptive adjustment item;

[0013] Step S40: Design a fuzzy rule base based on the characteristic dimensions of the high-order state vector to construct a fuzzy expert state analysis system for the energy storage power station; use the fuzzy expert state analysis system for the energy storage power station to perform a fusion analysis on the state point evolution trend of the battery packs in the battery energy storage power station, and output the health status and abnormal event information prediction results of the battery energy storage power station.

[0014] Furthermore, in step S10, the data cleaning includes removing abnormal data and repairing missing data; and preprocessing the historical data specifically includes:

[0015] The collected voltage , current ,capacity and internal resistance The original values of are compressed and smoothed respectively, and the processing formula is as follows:

[0016] ;

[0017] in, Indicates the value after compression and smoothing; represents the compression smoothing hyperparameter, according to the monitoring quantity The statistical characteristics of Obey the target distribution; Indicates voltage , current ,capacity and internal resistance The original value of Indicates voltage , current ,capacity and internal resistance The maximum value among the original values of ;

[0018] The collected temperature The original value of is processed as follows:

[0019] ;

[0020] in, Indicates the processed temperature sampling value; Indicates temperature The upper threshold value of The weight coefficient of the temperature anomaly factor is adjusted and optimized according to actual data and needs.

[0021] Furthermore, the specific process of step S20 includes:

[0022] S201: Extract the time series-frequency domain mixed features from the preprocessed data to obtain the time domain features of each monitoring quantity and frequency domain characteristics , through the weighted fusion of time domain features and frequency domain features, the multi-order feature vector is constructed ( , , , , ),in, represents the voltage eigenvector, represents the current eigenvector, represents the capacity feature vector, represents the internal resistance eigenvector, represents the temperature eigenvector;

[0023] The calculation formula of weighted fusion is as follows:

[0024] ;

[0025] in, Indicates specific monitored quantities, including voltage, current, capacity, internal resistance and temperature; and Represents the weighting coefficient, which adjusts the weight of each feature type according to actual needs.

[0026] Furthermore, the specific process of step S20 also includes:

[0027] S202: The multi-order feature vector ( , , , , ) Input the relationship fusion module to calculate the coupling parameters between the parameters in the multi-order feature vector. The coupling parameter calculation formula is:

[0028] ;

[0029] in, and Represents the multi-order eigenvector ( , , , , ) in two different parameters, including , , , , ; Representation parameter and parameters The coupling parameter between ; and Indicates the parameter elements, Indicates the number of elements; Represents the second-order nonlinear transformation function, the formula is ,in 、 and Represents global sensitivity hyperparameters, local nonlinear hyperparameters, and local exponential hyperparameters, which are adjusted and optimized according to actual data and requirements. and Able to dynamically adjust the sensitivity of battery parameter hysteresis effect, It is a logarithmic compensation intensity factor, which is preset as a fixed constant through calibration experiments and is used to adapt to different battery systems;

[0030] Based on the coupling parameter, calculate the first evaluation index vector , the calculation formula is as follows:

[0031] ;

[0032] in, Represents a multilayer perceptron with 5 input channels and 1 output channel; Indicates the parameter and parameters The coupling parameter value between them.

[0033] Furthermore, the specific process of step S20 also includes:

[0034] S203: First, through the linear transformation matrix 、 The first evaluation indicator vector The aging degree of the batteries in the battery pack of the battery energy storage station Map to the same dimensional space, and then concatenate into the second evaluation index vector according to the feature dimension , and finally the second evaluation index vector As the battery pack state evaluation vector.

[0035] Furthermore, the specific process of step S30 includes:

[0036] S301: Using state evaluation vectors at several time points Constructing the state evaluation space of the battery pack of the battery energy storage power station ,in represents the state evaluation vector at time t.

[0037] Furthermore, the specific process of step S30 also includes:

[0038] S302: The deep temporal feature convolutional network (DTFCN) is composed of three improved ResNet12 convolutional layers for feature extraction. The output of each convolutional layer then enters the corresponding activation function NLA-AF for nonlinear mapping processing. An LSTM network is immediately followed by the last ResNet12 in the DTFCN to further enhance the representation capability of long-term temporal features.

[0039] The NLA-AF is The activation function applied at each scale automatically selects the appropriate mapping according to the data fluctuation characteristics. The formula of the activation function is:

[0040] ;

[0041] in, For scale Time scale adjustment factor under ;

[0042] In the DTFCN, the state evaluation vector at each time point is extracted through multi-layer convolution operations. At the same time, the multi-scale nonlinear adaptive activation function NLA-AF is applied to the output of each layer for nonlinear mapping. The formula is:

[0043] ;

[0044] in, represents the state evaluation vector of the NLA-AF at time t The corresponding output; Indicates the first scales; and Represents the state evaluation vector at time t The corresponding convolutional layer weights and biases.

[0045] Furthermore, the specific process of step S30 also includes:

[0046] S303: Adaptive adjustment of local trajectory items through the following mechanism To update:

[0047] ;

[0048] in, Indicates the current The high-order state vector obtained by time step prediction; and Represents the prediction error correction hyperparameters and state transformation inertia hyperparameters, which are adjusted and optimized according to actual data and needs;

[0049] The high-order state vector Expressed as: , among which, the comprehensive health index Describe the overall health status of the battery pack; long-term trend characteristics Reflects the time trend of state evolution; abnormal behavior characteristics Nonlinear expression for sudden changes; dynamic coupling characteristics Quantitative indicators of complex relationships between monitored quantities.

[0050] Furthermore, the specific process of step S40 is as follows:

[0051] S401: Combine the high-order state vector Design fuzzy rule base based on the characteristic dimension;

[0052] S402: Perform fuzzy reasoning based on the rule base and membership function:

[0053] According to the rule base and membership function, the high-order state vector Fuzzification of eigenvalues;

[0054] Perform rule reasoning: Use fuzzy rules to reason about the evolution of all battery pack state points, and combine multiple high-order state vectors to generate the health status and abnormal risk of the battery energy storage power station;

[0055] Defuzzification: Use weighted average method to convert fuzzy results into specific numerical outputs;

[0056] S403: Output the evaluation results, including the overall health status of the battery energy storage power station and the abnormal event information prediction results.

[0057] The present invention also proposes a battery energy storage power station status assessment and intelligent perception system, which includes a data acquisition module, a data preprocessing module, a feature extraction module, a deep temporal feature convolutional network module and a fuzzy expert system analysis module.

[0058] The data acquisition module is used to collect monitoring data of each battery pack in the battery energy storage power station, including parameters such as voltage, current, capacity, internal resistance and temperature of the battery pack;

[0059] The data preprocessing module is used to clean and normalize the monitoring data to obtain preprocessed data;

[0060] The feature extraction module is used to calculate the multi-order feature vectors of each monitoring data based on the preprocessed data, and extract the complex coupling relationship between the monitoring data through the relationship fusion module to construct a state assessment vector for quantifying the operating status of the battery pack of the battery energy storage power station;

[0061] The deep temporal feature convolutional network module is used to construct a state evaluation space of the battery pack of the battery energy storage station based on the state evaluation vector, and to model the state point evolution trajectory of the battery pack of the battery energy storage station through a deep learning network within the state evaluation space;

[0062] The fuzzy expert system analysis module is used to perform a fusion analysis of the state point evolution trajectories of all battery packs in the battery energy storage power station, and based on fuzzy logic reasoning, predict and output the overall health status information of the battery energy storage power station, including health status, output power and abnormal event prediction results.

[0063] Compared with the prior art, the present invention has the following beneficial effects:

[0064] 1. Improving operational status awareness: This invention uses a multi-order feature vector construction method based on time-series-frequency domain hybrid feature extraction to comprehensively extract the time-series and frequency-domain characteristics of each monitored variable (voltage, current, capacity, internal resistance, and temperature) in a battery energy storage power station. Furthermore, a relationship fusion module is used to capture the complex coupling relationships between these monitored variables, thereby accurately quantifying the operational status of the battery packs in the battery energy storage power station.

[0065] 2. Enhanced health status assessment accuracy: By constructing a state assessment space and combining it with a deep temporal feature convolutional network, this invention can comprehensively capture the dynamic evolution of the state points of each battery pack in a battery energy storage power station, accurately assess the health status of the battery energy storage station, and significantly improve the accuracy and robustness of the state assessment.

[0066] 3. Promote intelligent operation and maintenance: By achieving intelligent perception, accurate status assessment, and operational trend prediction of battery energy storage power stations, this invention can significantly improve the intelligent operation and maintenance level of energy storage power stations, promote the application of the "condition-based maintenance" model, ensure the safe and efficient operation of energy storage power stations, and contribute to the development of new power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a flow chart of the battery energy storage power station status assessment and intelligent perception method of the present invention;

[0068] Figure 2This is a functional block diagram of the monitoring feature extraction module of the battery energy storage power station of the present invention;

[0069] Figure 3 This is a block diagram of the deep temporal feature convolutional network structure of the present invention. DETAILED DESCRIPTION

[0070] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit 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.

[0071] The present invention proposes a battery energy storage power station status assessment and intelligent perception method. The specific steps include: Step S10, collecting historical data of various monitoring quantities in the battery energy storage power station, including the voltage of each battery pack , current ,capacity , internal resistance and temperature , and perform data cleaning and preprocessing to obtain preprocessed data.

[0072] Specifically, Figure 1 The following is a flow chart of the battery energy storage power station status assessment and intelligent perception method. Figure 1 As shown, in step S10, the historical data is cleaned by a data processing module to remove "dirty" data and repair missing data; then the historical data is preprocessed to obtain the preprocessed data. The specific processing method is as follows:

[0073] The collected voltage , current ,capacity and internal resistance The original values of are compressed and smoothed respectively, and the processing formula is as follows:

[0074] ;

[0075] in, Indicates the value after compression and smoothing; represents the compression smoothing hyperparameter, according to the monitoring quantity (voltage / current / capacity / internal resistance) statistical characteristics (skewness, kurtosis) are adjusted to make Obey the target distribution (such as Gaussian or uniform distribution); Indicates voltage , current ,capacity and internal resistance The original value of represents the logarithmic function with base 10; Indicates voltage , current ,capacity and internal resistance The maximum value among the original values of ;

[0076] Hyperparameters were obtained through extensive experiments The empirical value range of is shown in the following table:

[0077] Table 1 Hyperparameters Example of an empirical value range for

[0078]

[0079] Preferably, in this embodiment, the hyperparameter According to the input data of each monitoring quantity, dynamic adjustment is performed within the range of the above table so that the voltage and current of the compressed and smoothed monitoring data obey Gaussian distribution, and the capacity and internal resistance obey uniform distribution.

[0080] The collected temperature The original value of is processed as follows:

[0081] ;

[0082] in, Indicates the processed temperature sampling value; Indicates temperature The upper threshold value of The weight coefficient representing the temperature anomaly factor is adjusted and optimized according to the actual data and requirements. The optimization process follows the common hyperparameter tuning paradigm in machine learning. In this embodiment, the initial search interval is determined to be [0.1, 5.0] based on prior experiments. When there are fewer abnormal practices in the temperature data, it is recommended to increase To enhance the capture of signals near the threshold; on the contrary, if the data noise is too much, it is necessary to reduce To suppress overfitting.

[0083] Collecting core monitoring data from battery energy storage power plants and performing data cleaning and preprocessing provides an accurate and reliable data foundation for subsequent analysis. Eliminating abnormal data prevents interference with assessment results and ensures data authenticity and validity. Cleaned and preprocessed data helps more accurately reflect the actual operating status of the battery, providing strong support for comprehensive quantification of battery operating status, accurate assessment of health status, and prediction of operating trends, enhancing the credibility of the entire assessment and perception process.

[0084] Figure 2This is a functional block diagram of a monitoring quantity feature extraction module 600 for a battery energy storage power station.

[0085] Step S20: Input the preprocessed data into a time-frequency domain hybrid feature extraction module to construct a multi-order feature vector of the monitoring quantity; extract the complex coupling relationship between the monitoring quantities in the multi-order feature vector through a relationship fusion module; generate a state assessment vector based on the complex coupling relationship to quantify the operating status of the battery pack of the battery energy storage power station.

[0086] The specific process of step S20 includes:

[0087] S201: Extract the time series-frequency domain mixed features from the preprocessed data to obtain the time domain features of each monitoring quantity and frequency domain characteristics , through the weighted fusion of time domain features and frequency domain features, the multi-order feature vector is constructed ( , , , , ),in, represents the voltage eigenvector, represents the current eigenvector, represents the capacity feature vector, represents the internal resistance eigenvector, represents the temperature eigenvector;

[0088] The calculation formula of weighted fusion is as follows:

[0089] ;

[0090] in, Indicates specific monitored quantities, including voltage, current, capacity, internal resistance and temperature; and Represents the weighting coefficient, and adjusts the weight of each feature type according to actual needs; normalized weight is used here: and Preferably, in this embodiment, the weighted coefficient calculation adopts the feature importance scoring method, firstly, the random forest or XGBoost model is used to calculate the time domain feature importance score. and frequency domain feature importance scores , and then distribute the weights proportionally:

[0091] , ;

[0092] The data of each monitoring quantity are fused to obtain the multi-order feature vector ( , , , , ).

[0093] S202: The multi-order feature vector ( , , , , ) Input the relationship fusion module to calculate the coupling parameters between the parameters in the multi-order feature vector. The coupling parameter calculation formula is:

[0094] ;

[0095] in, and Represents the multi-order eigenvector ( , , , , ) in two different parameters, including , , , , ; Representation parameter and parameters The coupling parameter between ; and Indicates the parameter elements, Indicates the number of elements; Represents the second-order nonlinear transformation function, the formula is ,in 、 and Represents global sensitivity hyperparameters, local nonlinear hyperparameters, and local exponential hyperparameters, which are adjusted and optimized according to actual data and requirements. and Able to dynamically adjust the sensitivity of battery parameter hysteresis effect, is the logarithmic compensation intensity factor, which is preset as a fixed constant through calibration experiments and is used to adapt to different battery systems; Indicates the absolute value operation;

[0096] Hyperparameters were obtained through extensive experiments 、 and The empirical value range of is shown in the following table:

[0097] Table 2 Hyperparameters 、 and Example of an empirical value range for

[0098]

[0099] Preferably, in this embodiment, the hyperparameter 、 and Determined by:

[0100] (a) Select the initial experience value based on the battery type, LFP battery =0.8, =2.5, =0.05;

[0101] (b) A joint optimization method with regularized gradient descent is used, and the objective function is the mean square error between the predicted coupling degree and the measured value;

[0102] (c) Use grid search method to conduct joint search within the range of the above table, and the optimal coupling degree corresponds to 、 and The value is the final parameter value.

[0103] Based on the coupling parameter, a first evaluation index vector is calculated. , the calculation formula is as follows:

[0104] ;

[0105] in, Represents a multilayer perceptron with 5 input channels and 1 output channel; Indicates the parameter and parameters The coupling parameter value between them.

[0106] S203: First, through the linear transformation matrix 、 The first evaluation indicator vector The aging degree of the batteries in the battery pack of the battery energy storage station Mapped to the same dimensional space. Preferably, in this embodiment, the linear transformation matrix 、 Generate it by following these steps:

[0107] (a) Matrix initialization based on Xavier distribution or pre-trained autoencoder;

[0108] (b) With the goal of minimizing the battery pack state assessment error, the matrix elements are jointly optimized using the gradient descent method;

[0109] (c) L2 regularization constraint is used to prevent matrix overfitting.

[0110] Then, the second evaluation index vector is formed by splicing the feature dimensions. , and finally the second evaluation index vector As the battery pack state evaluation vector.

[0111] Multi-level feature extraction and complex relationship fusion can fully tap into useful information from multidimensional monitoring variables. By comprehensively extracting the features of monitored variables such as voltage and current and integrating the complex relationships between them, this information can be mapped into a state assessment space, achieving comprehensive quantification of the battery's operating status. This not only provides a more detailed portrayal of the battery's state but also captures potential connections that are difficult to detect through single-parameter analysis. This provides a richer and more valuable quantitative basis for accurately assessing the operating status of battery energy storage power plants, improving the comprehensiveness and accuracy of assessments.

[0112] Figure 3 This is a block diagram of the deep temporal feature convolutional network structure.

[0113] Step S30: construct a state evaluation space based on the state evaluation vector ; In the state evaluation space In the embodiment, the state point evolution trajectory of the battery pack of the battery energy storage power station is modeled by a deep temporal feature convolutional network to obtain a high-order state vector in the state point evolution trajectory.

[0114] The specific process of step S30 includes:

[0115] S301: Using state evaluation vectors at several time points Constructing the state evaluation space of the battery pack of the battery energy storage power station ,in represents the state evaluation vector at time t.

[0116] S302: The Deep Temporal Feature Convolutional Network (DTFCN) consists of three improved ResNet12 convolutional layers for feature extraction. The output of each convolutional layer then enters the corresponding activation function NLA-AF for nonlinear mapping processing to adapt to the different fluctuation characteristics of time series data and better extract deep features. The last ResNet12 in the DTFCN is followed by an LSTM network to further enhance the representation of long-term time series features.

[0117] The NLA-AF is The activation function applied at each scale automatically selects the appropriate mapping according to the data fluctuation characteristics. The formula of the activation function is:

[0118] ;

[0119] in, For scale Time scale adjustment factor under ; represents a natural constant;

[0120] Preferably, in this embodiment According to the three convolutional layers of DTFCN, it is divided into three scales. The initial values of each scale are set to 0.5, 1.2, and 2.5, respectively. The dynamic adjustment strategy is to dynamically adjust according to the Hurst exponent H of the activation function input sequence to control the curvature of the output feature space. The specific formula is as follows:

[0121] ;

[0122] In the DTFCN, the state evaluation vector at each time point is extracted through multi-layer convolution operations. At the same time, the multi-scale nonlinear adaptive activation function NLA-AF is applied to the output of each layer for nonlinear mapping. The formula is:

[0123] ;

[0124] in, represents the state evaluation vector of the NLA-AF at time t The corresponding output; Indicates the first scales; and Represents the state evaluation vector at time t The corresponding convolutional layer weights and biases.

[0125] S303: Obtain a high-order state vector of the evolution trajectory of the state points of each battery pack in the battery energy storage station through the DTFCN prediction, and model the high-order state vector. The formula is:

[0126] ;

[0127] in, Indicates the The high-order state vector obtained by time step prediction; represents the DTFCN; Indicates the The state evaluation vector corresponding to the time step; Represents the local trajectory adaptive adjustment term, which is updated through the following mechanism:

[0128] ;

[0129] in, Indicates the current The high-order state vector obtained by time step prediction; and Represents the prediction error correction hyperparameters and state transformation inertia hyperparameters, which are adjusted and optimized according to actual data and needs; Characterizes the model's sensitivity to short-term fluctuations. The larger the value, the stronger the model's online update capability. Reflects the degree of system dependence on historical trend inertia. The larger the value, the better the trajectory smoothness. and Stability conditions must be met: and ;

[0130] Based on the empirical data of the prediction scenario, the hyperparameters 、 The initial value range of is shown in the following table:

[0131] Table 3 Hyperparameters 、 Example of the initial value range of

[0132]

[0133] Preferably, in this embodiment, the hyperparameter 、 Determined by:

[0134] (a) Select the initial value based on the forecast time scale. In this example, the short-term forecast is selected. =0.6, =0.1;

[0135] (b) Using rolling time window joint optimization algorithm to and Perform joint optimization, and the optimization formula is as follows:

[0136] , ;

[0137] in is the exponential decay weight; is a regularization hyperparameter, a non-negative real number, dynamically adjusted according to the prediction error, The initial value and dynamic adjustment rules are as follows:

[0138] If the data signal-to-noise ratio (SNR) is less than 10dB, The initial value is 0.5, otherwise it is 0.1;

[0139] when The prediction error When the 10 steps are raised continuously, the To enhance regularization, The dynamic update formula is: ,in =0.1 is the adjustment rate;

[0140] (c) Verification of the stability of the regulation process by Lyapunov exponent.

[0141] The high-order state vector Expressed as: , among which, the comprehensive health index Describe the overall health status of the battery pack; long-term trend characteristics Reflects the time trend of state evolution; abnormal behavior characteristics Nonlinear expression for sudden changes; dynamic coupling characteristics Quantitative indicators of complex relationships between monitored quantities.

[0142] Using a deep temporal feature convolutional network to model the evolution of battery pack state points effectively captures how battery state changes over time. Through multi-layer convolution operations and multi-scale nonlinear adaptive activation functions, deep features in time series can be extracted and adapted to the varying volatility of the data. Incorporating a long-short-term memory network enhances the representation of long-term features, enabling better prediction of future battery pack state changes. This helps identify potential battery issues in advance, providing a scientific basis for developing appropriate operation and maintenance strategies and ensuring the stable operation of battery energy storage power plants.

[0143] Step S40: Design a fuzzy rule base based on the characteristic dimensions of the high-order state vector to construct a fuzzy expert state analysis system for the energy storage power station; use the fuzzy expert state analysis system for the energy storage power station to perform a fusion analysis on the state point evolution trend of the battery packs in the battery energy storage power station, and output the health status and abnormal event information prediction results of the battery energy storage power station.

[0144] The specific process of step S40 is:

[0145] S401: Combine the high-order state vector The fuzzy rule base is designed based on the characteristic dimension of . The membership function is defined as follows:

[0146] Comprehensive health index, with membership functions of Healthy, Warning, and Fault.

[0147] For long-term trend changes, the membership functions are Stable, Rising, and Declining.

[0148] Abnormal behavior characteristics, the membership function is no anomaly (NoAnomaly), slight anomaly (MinorAnomaly), severe anomaly (SevereAnomaly).

[0149] Dynamic coupling characteristics, the membership functions are weak association (Weak), moderate association (Moderate), and strong association (Strong).

[0150] Fuzzy rule examples:

[0151] Rule 1: If the value is "Fault" and "Severe Abnormal", the health status is "Fault" and the abnormal risk is "High".

[0152] Rule 2: If it is "Declining" and "Strongly Correlated", the health status is "Warning" and the abnormal risk is "Medium".

[0153] Rule 3: If the value is “Healthy” and “No Abnormality”, the health status is “Healthy” and the abnormality risk is “Low”.

[0154] S402: Perform fuzzy reasoning based on the rule base and membership function. First, the high-order state vector The eigenvalues of the fuzzy matrix are fuzzy. For example, the membership of the comprehensive health index may be: Healthy: 0.8; Warning: 0.2; Fault: 0;

[0155] Then, rule reasoning is performed: fuzzy rules are used to reason about the evolution of all battery pack state points, and the health status and abnormal risk of the battery energy storage power station are generated by combining multiple high-order state vectors;

[0156] Finally, defuzzification is performed: using the weighted average method, the fuzzy results are converted into specific numerical outputs, including:

[0157] health status score (range 0-100, higher values indicate better health);

[0158] Abnormal risk score (range 0-100, higher values indicate greater risk).

[0159] S403: Output the evaluation results. The final output of the energy storage power station fuzzy expert state analysis system includes:

[0160] (1) Overall health status of the battery energy storage power station: Output health score (0-100), and divide the health status into: healthy (80-100); warning (50-79); fault (0-49).

[0161] (2) Possible abnormal event information: Combined with the abnormal risk score, the possible abnormal events and risk levels are output, for example:

[0162] Abnormal event 1: The comprehensive health index of a battery group decreases, and the long-term trend shows attenuation.

[0163] Abnormal event 2: The dynamic coupling characteristics of a certain battery group are abnormal, which may be due to temperature control or internal resistance problems.

[0164] The fuzzy expert state analysis system integrates and analyzes the evolution trends of battery pack status points, outputting the health status of the battery energy storage plant and information on possible abnormal events. Based on a fuzzy rule base and inference mechanism, it transforms complex battery status information into intuitive health scores and abnormality risk assessments. This enables operations and maintenance personnel to quickly understand the overall status of the energy storage plant, promptly identify potential anomalies, and conduct targeted maintenance and management, thereby improving the intelligent operation and maintenance of the battery energy storage plant and reducing operational risks.

[0165] The present invention also proposes a battery energy storage power station status assessment and intelligent perception system, which includes a data acquisition module, a data preprocessing module, a feature extraction module, a deep temporal feature convolutional network module and a fuzzy expert system analysis module.

[0166] The data acquisition module is used to collect monitoring data of each battery pack in the battery energy storage power station, including parameters such as voltage, current, capacity, internal resistance and temperature of the battery pack;

[0167] The data preprocessing module is used to clean and normalize the monitoring data to obtain preprocessed data;

[0168] The feature extraction module is used to calculate the multi-order feature vectors of each monitoring data based on the preprocessed data, and extract the complex coupling relationship between the monitoring data through the relationship fusion module to construct a state assessment vector for quantifying the operating status of the battery pack of the battery energy storage power station;

[0169] The deep temporal feature convolutional network module is used to construct a state evaluation space of the battery pack of the battery energy storage station based on the state evaluation vector, and to model the state point evolution trajectory of the battery pack of the battery energy storage station through a deep learning network within the state evaluation space;

[0170] The fuzzy expert system analysis module is used to perform a fusion analysis of the state point evolution trajectories of all battery packs in the battery energy storage power station, and based on fuzzy logic reasoning, predict and output the overall health status information of the battery energy storage power station, including health status, output power and abnormal event prediction results.

[0171] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A battery energy storage power station status assessment and intelligent perception method, characterized in that: include: Step S10: Collect historical data of various monitoring quantities in the battery energy storage power station, including the voltage of each battery pack , current ,capacity , internal resistance and temperature , and perform data cleaning and preprocessing; Step S20: Input the preprocessed data into a time-frequency domain hybrid feature extraction module to construct a multi-order feature vector of the monitoring quantity; extract the complex coupling relationship between the monitoring quantities in the multi-order feature vector through a relationship fusion module; generate a state assessment vector based on the complex coupling relationship to quantify the operating status of the battery pack of the battery energy storage power station; Step S30: construct a state evaluation space based on the state evaluation vector ; In the state evaluation space In the process, the state point evolution trajectory of the battery pack of the battery energy storage power station is modeled by a deep temporal feature convolutional network to obtain the high-order state vector in the state point evolution trajectory. The modeling formula is as follows: ； in, Indicates the The high-order state vector obtained by time step prediction; Represents the deep temporal feature convolutional network DTFCN; Indicates the The state evaluation vector corresponding to the time step; Represents the local trajectory adaptive adjustment item; Step S40: Design a fuzzy rule base based on the characteristic dimensions of the high-order state vector to construct a fuzzy expert state analysis system for the energy storage power station; use the fuzzy expert state analysis system for the energy storage power station to perform a fusion analysis on the state point evolution trend of the battery packs in the battery energy storage power station, and output the health status and abnormal event information prediction results of the battery energy storage power station.

2. The battery energy storage power station status assessment and intelligent perception method according to claim 1 is characterized in that: In step S10, the data cleaning includes removing abnormal data and repairing missing data; and preprocessing the historical data, specifically including: The collected voltage , current ,capacity and internal resistance The original values of are compressed and smoothed respectively, and the processing formula is as follows: ； in, Indicates the value after compression and smoothing; represents the compression smoothing hyperparameter, according to the monitoring quantity The statistical characteristics of Obey the target distribution; Indicates voltage , current ,capacity and internal resistance The original value of Indicates voltage , current ,capacity and internal resistance The maximum value among the original values of ; The collected temperature The original value of is processed as follows: ； in, Indicates the temperature sampling value after processing; Indicates temperature The upper threshold value of The weight coefficient of the temperature anomaly factor is adjusted and optimized according to actual data and needs.

3. The battery energy storage power station status assessment and intelligent perception method according to claim 1 is characterized in that: The specific process of step S20 includes: S201: Extract the time series-frequency domain mixed features from the preprocessed data to obtain the time domain features of each monitoring quantity and frequency domain characteristics , through the weighted fusion of time domain features and frequency domain features, the multi-order feature vector is constructed ( , , , , ),in, represents the voltage eigenvector, represents the current eigenvector, represents the capacity feature vector, represents the internal resistance eigenvector, represents the temperature eigenvector; The calculation formula of weighted fusion is as follows: ； in, Indicates specific monitored quantities, including voltage, current, capacity, internal resistance and temperature; and Represents the weighting coefficient, which adjusts the weight of each feature type according to actual needs.

4. The battery energy storage power station status assessment and intelligent perception method according to claim 3 is characterized in that: The specific process of step S20 also includes: S202: The multi-order feature vector ( , , , , ) Input the relationship fusion module to calculate the coupling parameters between the parameters in the multi-order feature vector. The coupling parameter calculation formula is: ； in, and Represents the multi-order eigenvector ( , , , , ) in two different parameters, including , , , , ; Representation parameter and parameters The coupling parameter between ; and Indicates the parameter elements, Indicates the number of elements; Represents the second-order nonlinear transformation function, the formula is ,in 、 and Represents global sensitivity hyperparameters, local nonlinear hyperparameters, and local exponential hyperparameters, which are adjusted and optimized according to actual data and requirements. and Able to dynamically adjust the sensitivity of battery parameter hysteresis effect, is the logarithmic compensation intensity factor, which is preset as a fixed constant through calibration experiments and is used to adapt to different battery systems; Based on the coupling parameter, calculate the first evaluation index vector , the calculation formula is as follows: ； in, Represents a multilayer perceptron with 5 input channels and 1 output channel; Indicates the parameter and parameters The coupling parameter value between them.

5. The battery energy storage power station status assessment and intelligent perception method according to claim 4 is characterized in that: The specific process of step S20 also includes: S203: First, through the linear transformation matrix 、 The first evaluation indicator vector The aging degree of the batteries in the battery pack of the battery energy storage station Map to the same dimensional space, and then concatenate into the second evaluation index vector according to the feature dimension , and finally the second evaluation index vector As the battery pack state evaluation vector.

6. The battery energy storage power station status assessment and intelligent perception method according to claim 5, characterized in that: The specific process of step S30 includes: S301: Using state evaluation vectors at several time points Constructing the state evaluation space of the battery pack of the battery energy storage power station ,in represents the state evaluation vector at time t.

7. The battery energy storage power station status assessment and intelligent perception method according to claim 6 is characterized in that: The specific process of step S30 also includes: S302: The deep temporal feature convolutional network (DTFCN) is composed of three improved ResNet12 convolutional layers for feature extraction. The output of each convolutional layer then enters the corresponding activation function NLA-AF for nonlinear mapping processing. An LSTM network is immediately followed by the last ResNet12 in the DTFCN to further enhance the representation capability of long-term temporal features. The NLA-AF is The activation function applied at each scale automatically selects the appropriate mapping according to the data fluctuation characteristics. The formula of the activation function is: ； in, For scale Time scale adjustment factor under ; In the DTFCN, the state evaluation vector at each time point is extracted through multi-layer convolution operations. At the same time, the multi-scale nonlinear adaptive activation function NLA-AF is applied to the output of each layer for nonlinear mapping. The formula is: ； in, represents the state evaluation vector of the NLA-AF at time t The corresponding output; Indicates the first scales; and Represents the state evaluation vector at time t The corresponding convolutional layer weights and biases.

8. The battery energy storage power station status assessment and intelligent perception method according to claim 7, characterized in that: The specific process of step S30 also includes: S303: Adaptive adjustment of local trajectory items through the following mechanism To update: ； in, Indicates the current The high-order state vector obtained by time step prediction; and Represents the prediction error correction hyperparameters and state transformation inertia hyperparameters, which are adjusted and optimized according to actual data and needs; The high-order state vector Expressed as: , among which, the comprehensive health index Describe the overall health status of the battery pack; long-term trend characteristics Reflects the time trend of state evolution; abnormal behavior characteristics Nonlinear expression for sudden changes; dynamic coupling characteristics Quantitative indicators of complex relationships between monitored quantities.

9. The battery energy storage power station status assessment and intelligent perception method according to claim 1, characterized in that: The specific process of step S40 is: S401: Combine the high-order state vector Design fuzzy rule base based on the characteristic dimension; S402: Perform fuzzy reasoning based on the rule base and membership function: According to the rule base and membership function, the high-order state vector Fuzzification of eigenvalues; Perform rule reasoning: Use fuzzy rules to reason about the evolution of all battery pack state points, and combine multiple high-order state vectors to generate the health status and abnormal risk of the battery energy storage power station; Defuzzification: Use weighted average method to convert fuzzy results into specific numerical outputs; S403: Output the evaluation results, including the overall health status of the battery energy storage power station and the abnormal event information prediction results.

10. A battery energy storage power station status assessment and intelligent perception system using the method according to any one of claims 1 to 9, comprising a data acquisition module, a data preprocessing module, a feature extraction module, a deep temporal feature convolutional network module, and a fuzzy expert system analysis module, characterized in that: The data acquisition module is used to collect monitoring data of each battery pack in the battery energy storage power station, including parameters such as voltage, current, capacity, internal resistance and temperature of the battery pack; The data preprocessing module is used to clean and normalize the monitoring data to obtain preprocessed data; The feature extraction module is used to calculate the multi-order feature vectors of each monitoring data based on the preprocessed data, and extract the complex coupling relationship between the monitoring data through the relationship fusion module to construct a state assessment vector for quantifying the operating status of the battery pack of the battery energy storage power station; The deep temporal feature convolutional network module is used to construct a state evaluation space of the battery pack of the battery energy storage station based on the state evaluation vector, and to model the state point evolution trajectory of the battery pack of the battery energy storage station through a deep learning network within the state evaluation space; The fuzzy expert system analysis module is used to perform a fusion analysis of the state point evolution trajectories of all battery packs in the battery energy storage power station, and based on fuzzy logic reasoning, predict and output the overall health status information of the battery energy storage power station, including health status, output power and abnormal event prediction results.

Images