Energy storage system health state assessment method and system

Abstract

The invention discloses an energy storage system health state assessment method and system, mainly relates to the technical field of health state assessment, and is used for solving the problems that in the prior art, key degradation signals on a frequency domain cannot be identified, an entropy change rule of non-stable change cannot be reflected, and a deep neural network is insufficient in perception of a degradation region. Comprising the following steps: acquiring original monitoring data of the energy storage system in real time, calculating to obtain normalized data after time domain and frequency domain processing, and calculating characteristics of the enhanced monitoring data; constructing a residual attention deep neural network, and autonomously adjusting the current adaptive learning rate; dynamic focus loss is obtained through calculation, and a total loss function is calculated; obtaining a trained residual attention deep neural network; and acquiring current monitoring data, calculating features of the enhanced current monitoring data, and further performing energy storage system health state assessment by using the trained residual attention deep neural network.

Description

Technical Field

[0001] The present application relates to the technical field of health status assessment, and in particular to a method and system for assessing the health status of an energy storage system. Background Art

[0002] With the widespread adoption of new energy technologies, energy storage systems, especially lithium-ion battery storage devices, have played a key role in grid peak regulation, renewable energy integration, and electric transportation. However, over the long term, energy storage systems often experience performance degradation, such as capacity fade, increased internal resistance, and temperature anomalies.

[0003] The existing solution is to collect the working parameters of each working energy storage battery pack during its working process in real time, and collect the environmental parameters of the environment in which each working energy storage battery pack is located in real time, and calculate the fault diagnosis coefficient of each working energy storage battery pack based on the working parameters and environmental parameters of each working energy storage battery pack, and perform fault diagnosis on each working energy storage battery pack based on the fault diagnosis coefficient of each working energy storage battery pack. For the working energy storage battery pack with a fault, a spare energy storage battery pack is selected for replacement based on the rated capacity, charging efficiency and discharging efficiency of the energy storage battery pack.

[0004] However, the above technical solutions have the following issues that need to be further addressed: Conventional Min-Max or Z-score normalization cannot identify key degradation signals in the frequency domain, and is particularly insensitive to frequency variations with small amplitudes, resulting in indistinguishable degradation states. Traditional sliding window statistical features cannot reflect the entropy change patterns of non-stationary changes, are insufficiently capable of identifying abnormal fluctuations or early degradation trends, and exhibit significant lags. Conventional deep neural networks lack awareness of degraded regions and are unable to dynamically assign feature weights, which can easily lead to redundant feature propagation or the submersion of key information. Summary of the Invention

[0005] In view of the above-mentioned deficiencies in the prior art, the present application provides a method and system for evaluating the health status of an energy storage system to solve the above-mentioned technical problems.

[0006] In a first aspect, the present application provides a method for assessing the health status of an energy storage system, the method comprising:

[0007] Collect the original monitoring data of the energy storage system in real time and store the original monitoring data according to the preset time window;

[0008] The original monitoring data within the preset time window is processed in a unified time scale to obtain a time normalization result, and the original monitoring data is processed in the frequency domain to obtain a frequency domain compression result; the normalized data after time domain and frequency domain processing is calculated based on the time normalization result and the frequency domain compression result;

[0009] Based on the normalized data after time domain and frequency domain processing, the fluctuation characteristics of the time series are modeled in combination with the fluctuation entropy mechanism to calculate the characteristics of the enhanced monitoring data;

[0010] Construct a residual attention deep neural network consisting of an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer stacked in sequence; by monitoring the variance changes of the parameter gradients in each round of residual attention deep neural network training, the current adaptive learning rate is automatically adjusted;

[0011] Through the residual attention deep neural network, the health status probability of each health status category is output; according to the health status probability of each health status category and the preset loss weight corresponding to each category, the dynamic focus loss is calculated;

[0012] Obtain dynamic focus loss, temporal continuity loss, and attention sparse regularization loss to calculate the total loss function; based on the total loss function and adaptive learning rate, update the iterative trainable parameters; when the preset stop iteration condition is met, obtain the trained residual attention deep neural network;

[0013] The current monitoring data is obtained, the features of the enhanced current monitoring data are calculated, and then the trained residual attention deep neural network is used to evaluate the health status of the energy storage system.

[0014] In one implementation of the present application, real-time collection of raw monitoring data of the energy storage system specifically includes:

[0015] Real-time collection of raw monitoring data through a sensor network deployed on key components of the energy storage system;

[0016] The data sources of the original monitoring data include the voltage sensor, current sensor, temperature sensor of the battery cell / module, and the charge and discharge status information recorded by the battery management system;

[0017] The original monitoring data is collected using synchronous high-frequency sampling above 1 kHz.

[0018] In one implementation of the present application, the original monitoring data within a preset time window is uniformly processed in terms of time scale to obtain a time normalization result, and the original monitoring data is processed in the frequency domain to obtain a frequency domain compression result; the normalized data after time domain and frequency domain processing is calculated based on the time normalization result and the frequency domain compression result, specifically including:

[0019] In the time dimension, the mean and standard deviation of the original monitoring data within the preset time window are calculated, and based on this, the data at the current moment is centered and scaled to unify the time scale and obtain the time normalization result;

[0020] The time normalization result is mapped to the frequency domain space through the frequency domain transformation method to obtain the spectrum amplitude; the spectrum amplitude is compressed to the preset amplitude dynamic range using the logarithmic function;

[0021] The time normalization result is multiplied element by element by the frequency domain compression result to obtain the normalized data after time domain and frequency domain processing.

[0022] In one implementation of the present application, based on the normalized data after time domain and frequency domain processing, the fluctuation characteristics of the time series are modeled in combination with the fluctuation entropy mechanism, and the enhanced monitoring data features are calculated, specifically including:

[0023] Extracting normalized data at different time points through a sliding time window and calculating the distance between the extracted data and several pre-generated center vectors;

[0024] According to the distance, the exponential mapping is used to construct the probability that the normalized data belongs to each center, and the fluctuation entropy vector corresponding to the normalized data is calculated based on the probability of each center;

[0025] The normalized data is concatenated with the corresponding fluctuation entropy vector to obtain the enhanced monitoring data features.

[0026] In one implementation of the present application, a residual attention deep neural network is constructed by stacking an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer in sequence, specifically including:

[0027] The input layer receives the features of the enhanced monitoring data and maps the features of the enhanced monitoring data to the initial high-dimensional feature space of the network;

[0028] The residual path in the residual attention module extracts preset common features from the features of the enhanced monitoring data in the high-dimensional feature space, and the attention path dynamically adjusts the feature weights; through the preset common features, the features of the enhanced monitoring data, the self-attention weights, and the Sigmoid activation function, the features processed by the residual attention are obtained;

[0029] The hybrid regularization layer randomly discards a preset part of the residual attention module output with a preset probability to obtain the first feature, and adopts the spectral normalization mechanism to perform spectral normalization on the residual attention module output to obtain the second feature; the feature after the hybrid regularization of the first feature and the second feature is obtained;

[0030] The state output layer obtains the basic probability of each health state category through the conventional classification function;

[0031] Based on the number differences between each health state category, a state transition matrix is constructed;

[0032] The health state probability is calculated based on the basic probability, the features after hybrid regularization, and the state transition matrix.

[0033] In one implementation of the present application, the current adaptive learning rate is adjusted autonomously by monitoring the variance change of the parameter gradient in each round of residual attention deep neural network training, specifically including:

[0034] Calculate the ratio of the gradient variance of the current round to the gradient variance of the previous round; where the current round is greater than 1;

[0035] If the current gradient variance is large, the adaptive learning rate is relaxed; otherwise, the adaptive learning rate is gradually reduced; the relaxation / reduction of the learning rate is achieved through an exponential decay mechanism.

[0036] In one implementation of the present application, the dynamic focus loss is calculated based on the health state probability of each health state category and the preset loss weight corresponding to each category, specifically including:

[0037] By formula:

[0038] , calculate the dynamic focusing parameters of category c ;in, represents the number of samples in the cth category, c∈[1,C], C represents the total number of health status categories, represents the median number of samples in each category;

[0039] By formula:

[0040] , calculate the loss weight of category c ;in, represents the total number of samples, Indicates a preset constant;

[0041] By formula:

[0042] , calculate the dynamic focus loss ;in, Indicates the health status probability of the category;

[0043] In one implementation of the present application, dynamic focus loss, temporal continuity loss, and attention sparse regularization loss are obtained, and the total loss function is calculated, specifically including:

[0044] Obtain the corresponding weight values of dynamic focus loss, temporal continuity loss, and attention sparse regularization loss, and then weight them to obtain the total loss function.

[0045] In one implementation of the present application, based on the total loss function and the adaptive learning rate, the iterative trainable parameters are updated, specifically including:

[0046] Calculate the gradient of the trainable parameters using the total loss function and the trainable parameters of the current round;

[0047] Adopt gradient clipping to clip the gradient of trainable parameters;

[0048] Use the clipped gradient and adaptive learning rate to update the momentum term, and add a preset weight decay term to suppress overfitting;

[0049] Generate the trainable parameters for the next iteration based on the current momentum term, the updated momentum term, the adaptive learning rate, the current round of trainable parameters, and the gradient of the pruned trainable parameters.

[0050] In a second aspect, the present application provides an energy storage system health status assessment system, the system comprising:

[0051] The acquisition module is used to collect the original monitoring data of the energy storage system in real time and store the original monitoring data according to the preset time window;

[0052] The normalization calculation module is used to perform time-scale uniform processing on the original monitoring data within the preset time window to obtain a time normalization result, and perform frequency domain processing on the original monitoring data to obtain a frequency domain compression result; the normalized data after time domain and frequency domain processing is calculated based on the time normalization result and the frequency domain compression result;

[0053] The enhanced computing module is used to model the fluctuation characteristics of the time series based on the normalized data processed in the time and frequency domains, combined with the fluctuation entropy mechanism, and calculate the characteristics of the enhanced monitoring data;

[0054] The learning rate adjustment module is used to construct a residual attention deep neural network composed of an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer. By monitoring the variance changes of the parameter gradients in each round of residual attention deep neural network training, the current adaptive learning rate is automatically adjusted.

[0055] The loss calculation module is used to output the health status probability of each health status category through the residual attention deep neural network; calculate the dynamic focus loss based on the health status probability of each health status category and the preset loss weight corresponding to each category; obtain the dynamic focus loss, temporal continuity loss, and attention sparse regularization loss to calculate the total loss function;

[0056] The iterative training module is used to update the iterative trainable parameters based on the total loss function and adaptive learning rate; when the preset stopping condition is met, the trained residual attention deep neural network is obtained;

[0057] The real-time evaluation module is used to obtain current monitoring data, calculate the enhanced features of the current monitoring data, and then use the trained residual attention deep neural network to evaluate the health status of the energy storage system.

[0058] Those skilled in the art will appreciate that this application has at least the following beneficial effects:

[0059] 1. Combine time normalization with frequency domain logarithmic compression, and multiply the time normalization result with the frequency domain compression result element by element to obtain normalized data after time domain and frequency domain processing, so as to achieve synchronous normalization of time domain and frequency domain, eliminate non-stationary noise interference, retain key frequency characteristics, achieve suppression of non-stationary noise and enhancement of frequency degradation characteristics, and break through the expression limitations of single domain normalization methods.

[0060] 2. Construct probability distribution through dynamic clustering and exponential mapping, calculate the fluctuation entropy vector and splice it with the original normalized features to enhance the modeling ability of uncertain fluctuation patterns in the degradation process.

[0061] 3. Utilizing the residual structure and attention mechanism of the residual attention module, through skip connections and dynamic weight adjustment, we can strengthen degradation-related features, suppress redundant interference, and improve the model's ability to recognize degradation-sensitive areas.

[0062] 4. In the output layer, the state transfer matrix is combined to simulate the evolution law of the health state, and the focal loss is adjusted by the number of category samples to ensure that the classification results have temporal continuity and category balance. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] In order to more clearly illustrate the technical solution of the present invention, the following is a brief introduction to the drawings required for the description. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0064] Figure 1 This is a flow chart of a method for evaluating the health status of an energy storage system provided in an embodiment of the present application.

[0065] Figure 2 This is a diagram showing the influence of different normalization methods on the discrimination of health status features provided in an embodiment of the present application.

[0066] Figure 3 This is a diagram showing the impact of feature enhancement on early degradation detection capability provided by an embodiment of the present application.

[0067] Figure 4This is a training effect diagram of an adaptive learning rate optimization strategy provided in an embodiment of the present application.

[0068] Figure 5 This is a schematic diagram of the classification accuracy of different models under various health conditions provided in an embodiment of the present application.

[0069] Figure 6 This is an internal structure diagram of an energy storage system health status assessment system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0070] It should be understood by those skilled in the art that the embodiments described below are merely preferred embodiments of the present disclosure and do not imply that the present disclosure can only be implemented through these preferred embodiments. These preferred embodiments are merely intended to explain the technical principles of the present disclosure and are not intended to limit the scope of protection of the present disclosure. Based on the preferred embodiments provided by the present disclosure, all other embodiments obtained by those skilled in the art without creative effort should still fall within the scope of protection of the present disclosure.

[0071] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0072] The technical solutions proposed in the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0073] The embodiment provides a method for evaluating the health status of an energy storage system. Figure 1 As shown, the method provided in the embodiment of the present application mainly includes the following steps:

[0074] Step 110: Collect the original monitoring data of the energy storage system in real time, and store the original monitoring data according to a preset time window.

[0075] It should be noted that the raw monitoring data is collected in real time through the sensor network deployed on the key components of the energy storage system;

[0076] Data sources include the voltage sensor, current sensor, and temperature sensor of the battery cell / module, as well as the charge and discharge status information recorded by the battery management system. The current sensor can be a Hall effect sensor, and the temperature sensor can be a thermocouple sensor.

[0077] The acquisition method uses synchronous high-frequency sampling above 1 kHz to ensure the capture of transient fluctuation characteristics of parameters such as current.

[0078] In one embodiment, taking current data as an example, a current sensor connected in series in the battery circuit continuously collects real-time current sequences during the charging and discharging process, including steady-state operating current, transient impact current, and abnormal current.

[0079] At the same time, the collected data is categorized and manually labeled based on degradation indicators such as battery capacity decay and internal resistance growth. The health status is divided into five categories: "normal," "mild degradation," "moderate degradation," "severe degradation," and "fault." The labeling information comes from the calibration results of laboratory accelerated aging tests and fault diagnosis records in actual operation and maintenance.

[0080] All raw monitoring data are stored according to preset time windows to provide basic input for subsequent time-frequency normalization processing.

[0081] Step 120: Perform time-scale uniform processing on the original monitoring data within the preset time window to obtain a time normalization result, perform frequency domain processing on the original monitoring data to obtain a frequency domain compression result; and calculate the normalized data after time domain and frequency domain processing using the time normalization result and the frequency domain compression result.

[0082] Specifically, for the original sensor data, the mean and standard deviation within the preset time window are first calculated in the time dimension, and based on this, the data at the current moment is centered and scaled to unify its time scale. Then, the original data is mapped to the frequency domain space through the frequency domain transformation method to obtain the spectrum amplitude, and then the logarithmic function is used to compress its amplitude dynamic range. Then, the time normalization result and the frequency domain compression result are multiplied element by element to obtain the normalized data after time domain and frequency domain processing, so as to achieve synchronous normalization in the time domain and frequency domain, eliminating non-stationary noise interference and retaining key frequency characteristics, which can be expressed as:

[0083] ,

[0084] Where, is the normalized data after time domain and frequency domain processing at time t;

[0085] is the original sensor data vector at the tth moment, such as the current data vector collected by the original current sensor at the tth moment;

[0086] is the time domain mean, and the calculation method is expressed as ;

[0087] is the size of the data time window, representing the time series length of the data;

[0088] The first The raw sensor data vector at each time point;

[0089] is the time domain standard deviation, which is calculated as ;

[0090] It is a logarithmic function, and the default base is 10;

[0091] represents the fast Fourier transform, which converts the time domain signal into the frequency domain;

[0092] Represents the Hadamard product.

[0093] It should be noted that the energy storage sensor data is affected by working condition fluctuations, and the time domain statistics need to unify the dynamic range to suppress transient noise, and the time domain mean and time domain standard deviation By calculating the mean and standard deviation of the data in the sliding window, the time domain amplitude offset and scale difference are eliminated.

[0094] It should also be noted that battery aging causes the characteristic frequency amplitude to attenuate, such as the internal resistance characteristic frequency band. This application uses the logarithmic function term Amplify weak amplitude signals, take the logarithm of the spectrum amplitude after Fourier transform, compress the frequency domain energy in the high dynamic range, and retain the key frequency domain features related to degradation.

[0095] It should also be noted that the normalized data after time domain and frequency domain processing at time t In the calculation process, the conventional simple splicing method is not used. Instead, the time domain and frequency domain characteristics are coupled through Hadamard product calculation, while suppressing the time domain non-stationary noise and enhancing the frequency domain attenuation characteristics, solving the problem that conventional normalization ignores the frequency domain energy distribution.

[0096] In one embodiment, Figure 2As shown in the figure, in order to evaluate the impact of different normalization methods on the ability to distinguish the health status characteristics of the energy storage system, the three methods of traditional minimum and maximum normalization, standardized normalization (Z-score normalization) and the time-frequency dual-domain normalization proposed in this application are compared. The box plot results show that the traditional normalization method has significant overlap in the "mild degradation" to "severe degradation" state range, and the eigenvalue distribution range is wide and the boundaries are fuzzy, indicating that it is difficult to effectively distinguish adjacent degradation states. The eigenvalue boxes of the method in this application are compact under each health state, and the median position rises in a step-like manner. In particular, there is a clear gap between the "severe degradation" and "fault" states, indicating that the time-frequency dual-domain processing mechanism of this application uses a time domain sliding window to eliminate transient noise interference, frequency domain logarithmic compression to enhance weak degradation characteristics, and Hadamard product fusion to achieve synergistic enhancement of time-frequency characteristics, so that the characteristic distributions of different degradation states are essentially separated, laying the foundation for subsequent accurate classification.

[0097] Step 130: Based on the normalized data processed in the time domain and frequency domain, the fluctuation characteristics of the time series are modeled in combination with the fluctuation entropy mechanism, and the characteristics of the enhanced monitoring data are calculated.

[0098] This step can be specifically as follows: based on the normalized data after time domain and frequency domain processing, the fluctuation characteristics of the time series are modeled in combination with the fluctuation entropy mechanism. Specifically, first, data at different time points are extracted through a sliding time window, and the distance between the data and multiple pre-generated center vectors is calculated. Then, based on these distances, the probability distribution of the data belonging to each center is constructed using exponential mapping. Then, based on the probability distribution, its entropy value is calculated, which is expressed as the uncertainty intensity of the fluctuation. Then, the original normalized data is spliced with the corresponding fluctuation entropy vector to obtain the enhanced feature to enhance the differential expression related to the degradation trend in the data, which is expressed as:

[0099] ,

[0100] ,

[0101] ,

[0102] Where, For the enhanced features (features of the enhanced monitoring data), the original data is enhanced by combining them with the fluctuation entropy;

[0103] It is the normalized data after time domain and frequency domain processing;

[0104] For splicing operation;

[0105] It is the fluctuation entropy vector, which quantifies the uncertainty of fluctuation based on the entropy value. The increase of entropy indicates abnormal fluctuation, which can directly correlate to the degradation of health status.

[0106] For the Fluctuation entropy of the class;

[0107] is the time window length;

[0108] The data belongs to class probability;

[0109] represents a logarithmic function with a natural constant as the base;

[0110] represents an exponential function with a natural constant as its base;

[0111] is the fluctuation sensitivity factor, which controls the sensitivity of the probability distribution to the distance, for example, it is set to 0.5;

[0112] For the The cluster centers of the classes are generated by DBSCAN clustering to characterize typical fluctuation patterns and avoid the deviation caused by pre-set fixed centers;

[0113] is the number of dynamic cluster centers, determined by the DBSCAN clustering method.

[0114] It should be noted that Indicates the The fluctuation entropy of the class is the entropy value calculated based on the distance between the input data and the dynamic cluster center. is the entire fluctuation entropy vector, which contains the fluctuation entropy values of all categories.

[0115] It should be noted that the enhanced features In the calculation process, the conventional simple splicing method is not used. Instead, the entropy vector is used as a new feature to be spliced with the original data through splicing operations. The entropy vector provides global statistical characteristics of the degradation trend, which complements the local time series characteristics and jointly enhances the model's sensitivity to early weak degradation.

[0116] In one embodiment, Figure 3As shown, to analyze the effect of the fluctuating entropy feature on improving early degradation detection capabilities, a simulated accelerated degradation experiment was conducted, assuming that the battery would be unusable after 1,000 cycles. The comparison curve between the number of battery cycles and the degree of detected degradation shows that in the early degradation stage, the detection value of the method without the fluctuating entropy feature enhancement is significantly lower than the actual degradation curve, showing a serious lag, while the detection trajectory of the method applied closely matches the actual degradation process. By the severe degradation stage, the difference between the two methods has narrowed. The experimental results show that by calculating the probability distribution entropy value of the data point and the dynamic cluster center, the fluctuation uncertainty of the sensor data is quantified. The splicing and fusion of the entropy vector and the original features enables the model to simultaneously perceive local details and judge global fluctuation trends, significantly improving the model's sensitivity to weak early degradation signals.

[0117] Step 140: construct a residual attention deep neural network consisting of an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer stacked in sequence; by monitoring the variance change of the parameter gradient in each round during the training of the residual attention deep neural network, autonomously adjust the current adaptive learning rate.

[0118] Among them, the residual attention deep neural network is constructed by stacking the input layer, several residual attention modules, hybrid regularization layer and state output layer in sequence, specifically including:

[0119] The input layer receives the features of the enhanced monitoring data and maps the features of the enhanced monitoring data to the initial high-dimensional feature space of the network;

[0120] The residual path in the residual attention module extracts preset common features from the features of the enhanced monitoring data in the high-dimensional feature space, and the attention path dynamically adjusts the feature weights; through the preset common features, the features of the enhanced monitoring data, the self-attention weights, and the Sigmoid activation function, the features processed by the residual attention are obtained;

[0121] The hybrid regularization layer randomly discards a preset part of the residual attention module output with a preset probability to obtain the first feature, and adopts the spectral normalization mechanism to perform spectral normalization on the residual attention module output to obtain the second feature; the feature after the hybrid regularization of the first feature and the second feature is obtained;

[0122] The state output layer obtains the basic probability of each health state category through the conventional classification function;

[0123] Based on the number differences between each health state category, a state transition matrix is constructed;

[0124] The health state probability is calculated based on the basic probability, the features after hybrid regularization, and the state transition matrix.

[0125] To further illustrate, conventional deep neural networks are insufficient in modeling long-term dependencies and are prone to ignoring differences in sensor degradation contributions. This application uses an attention residual module in the deep neural network structure. For the input features of each layer, its feature representation is first extracted through a residual path, which includes two layers of fully connected networks and activation functions. At the same time, a set of attention weights is used to calculate the importance of different parts of the input features, and nonlinear mapping is performed through the activation function. Then, the attention-weighted features are element-wise multiplied by the residual path results and jump-connected with the original input to obtain features processed by residual attention, so as to enhance the propagation effect of important features and suppress redundant interference of irrelevant features, which is expressed as:

[0126] ;

[0127] ;

[0128] Where, is the feature after residual attention processing;

[0129] is the input feature of the residual attention deep neural network. For the first layer input of the residual attention deep neural network, That is ;

[0130] is the residual path, containing two fully connected layers;

[0131] is the Sigmoid activation function;

[0132] is the self-attention weight;

[0133] is the ReLU activation function;

[0134] is the first weight matrix, which is a trainable parameter;

[0135] is the first bias term, which is a trainable parameter;

[0136] is the second bias term, which is a trainable parameter;

[0137] is the second weight matrix, which is a trainable parameter.

[0138] Since small sample training easily leads to overfitting, the Dropout of conventional deep neural networks easily destroys the temporal correlation. In the deep network training process, this application adopts a hybrid approach to suppress the overfitting risk. Specifically, first, some residual path outputs are randomly discarded with a preset probability to avoid the model from relying on fixed features, thereby enhancing the generalization ability. Then, a spectral normalization mechanism is used to constrain the amplitude of the weight parameters in the residual module to prevent some weights from abnormally growing due to overfitting during the training process. The final feature after hybrid regularization is obtained by adding these two parts, taking into account both feature stability and training robustness, and is expressed as:

[0139] ;

[0140] ;

[0141] Where, is the feature after hybrid regularization processing;

[0142] Indicates randomly discarding residual paths with a preset probability, for example, the preset probability is 5%;

[0143] is the spectrum constraint strength, for example, set to 0.01;

[0144] is the spectral normalization constraint;

[0145] is the L2 norm;

[0146] The residual attention deep neural network layer weight matrix;

[0147] represents the maximum value function;

[0148] Since conventional deep neural networks use Softmax to output classification probabilities and ignore the transition probabilities between states, this application uses a state transition matrix to perform Markov modeling on the classification results at the output end. Specifically, the output of the last layer of the network obtains the basic probability of each health state through a conventional classification function. Then, based on the numbering differences between the states, a state transition matrix is constructed to simulate the transition trend between adjacent states. The final output of the model is the product of the basic classification result and the state transition matrix, so that the prediction result has smooth continuity in the time series, which is expressed as:

[0149] ;

[0150] ;

[0151] Where, is the Softmax function; Represents the weight matrix of the last layer of the network, Linear mapping to health state category space;

[0152] is the health status probability vector after being processed by the Markov output layer, and the classification category is determined by the category corresponding to the maximum probability value in the health status probability vector;

[0153] The state transition matrix is assigned transition probability based on the state number distance. In view of the continuity of the energy storage system monitoring data, the adjacent state transition probability is high and the jump state probability is low, forcing the prediction result to conform to the degenerate continuity.

[0154] is the transition probability from the i-th health state to the j-th health state;

[0155] is the transfer attenuation factor, for example, set to 0.5;

[0156] represents the state number difference between the i-th state and the j-th state;

[0157] represents the state number difference between the i-th state and the k-th state;

[0158] C is the classification category of the energy storage system health status.

[0159] In this step, the current adaptive learning rate is adjusted autonomously by monitoring the variance change of the parameter gradient in each round of residual attention deep neural network training, specifically including:

[0160] Calculate the ratio of the gradient variance of the current round to the gradient variance of the previous round; where the current round is greater than 1;

[0161] If the current gradient variance is large, the adaptive learning rate is relaxed; otherwise, the adaptive learning rate is gradually reduced; the relaxation / reduction of the learning rate is achieved through an exponential decay mechanism.

[0162] Specifically, conventional deep neural networks usually use a fixed learning rate, which is difficult to adapt to the staged convergence characteristics of health status classification. The present application monitors the variance changes of the parameter gradients in each round during network training and autonomously adjusts the current learning rate. Specifically, the ratio of the gradient variance of the current round to the gradient variance of the previous round is calculated. If the current gradient variance is large, the learning rate is relaxed to jump out of the local optimum. Otherwise, the learning rate is gradually reduced to promote convergence stability. The change of the learning rate is achieved through an exponential decay mechanism to control its change rate, which is expressed as:

[0163] ;

[0164] ;

[0165] Where, e is the training round;

[0166] is the adaptive learning rate of the e-th round;

[0167] is the initial learning rate, for example, set to 0.001;

[0168] is the learning rate attenuation coefficient, for example, set to 0.1;

[0169] is the parameter gradient vector of the e-th round;

[0170] For the wheel parameter gradient vector;

[0171] is the variance calculation function;

[0172] are the trainable parameters of the residual attention deep neural network;

[0173] Indicates the parameter quantity;

[0174] For the The gradient of the trainable parameters;

[0175] For the The gradient of a trainable parameter.

[0176] It should be noted that the degradation of health status occurs in stages. Characterizes the gradient variance and is used to monitor the stability of parameter updates. When the gradient variance is large, increasing the learning rate will help escape the local optimum.

[0177] In one embodiment, Figure 4As shown in the figure, the training process of the fixed learning rate and adaptive learning rate strategies is compared, and a double vertical axis line chart is used to simultaneously display the change in loss value and the learning rate adjustment trajectory. The fixed learning rate strategy falls into a plateau in the middle term, and the loss value stagnates at a high level. The adaptive strategy breaks through the platform by increasing the learning rate in the same round, and gradually converges to a lower loss level in the later stage. The learning rate curve shows that the adaptive mechanism maintains a high learning rate to accelerate convergence in the early stage of training, dynamically adjusts the learning rate according to the gradient variance in the middle term to jump out of the local optimum, and automatically decays in the final stage to ensure stable convergence. In the training of the health status assessment model, it is necessary to quickly approach the global optimal solution and avoid oscillation in the later stage, which significantly improves the model convergence efficiency and generalization ability.

[0178] Step 150: Output the health status probability of each health status category through the residual attention deep neural network; calculate the dynamic focus loss based on the health status probability of each health status category and the preset loss weight corresponding to each category, obtain the dynamic focus loss, temporal continuity loss, and attention sparse regularization loss, and calculate the total loss function.

[0179] Specifically, to avoid the majority class-dominated loss caused by sample imbalance, this application calculates the number of samples for each category and constructs corresponding weights based on this value to ensure that rare categories occupy a higher proportion in the loss function. At the same time, the focus factor of each category is dynamically adjusted according to the number of samples, so that prediction errors of rare categories are more severely penalized. The dynamic focus loss function can effectively improve the model's ability to recognize difficult samples by focusing on the probability output of each category, which is expressed as:

[0180] ;

[0181] ;

[0182] ;

[0183] Where, is the dynamic focus loss function;

[0184] is the loss weight of category c;

[0185] is the classification probability of the cth class;

[0186] Dynamically focus on the parameters and adjust the loss weight of difficult samples;

[0187] is the total number of samples;

[0188] is the number of samples in the cth category;

[0189] For preset constants, for example, ;

[0190] Represents the median of the number of samples in each category.

[0191] Calculate the total loss function:

[0192] In order to collaboratively optimize the accuracy of health status classification and the consistency of time series prediction, this application adopts a multi-task coupling loss mechanism. The total loss function consists of three parts: dynamic focus classification loss function Solve the problem of sample imbalance; time series continuity loss function Constrain the prediction results of adjacent time windows to conform to the Markov transition law; attention sparse regularization loss function Suppress excessive dispersion of attention weights and improve feature selection focus.

[0193] Specifically, the total loss function is defined as a weighted summation:

[0194]

[0195] Where, is the total loss function;

[0196] is the first loss weight, for example, set to 0.6;

[0197] is the second loss weight, for example, set to 0.2;

[0198] is the third loss weight, for example, set to 0.2.

[0199] The temporal continuity loss function is obtained by calculating the KL divergence of the state probabilities of adjacent time windows. The Markov transfer matrix is back-propagated to the training process to force the network to learn the continuity of state evolution. The calculation method is expressed as:

[0200] ;

[0201] Where, is the KL divergence calculation function;

[0202] Indicates the State probability vector of the time window;

[0203] is the state transition matrix.

[0204] The attention sparse regularization loss function calculates the negative entropy of the attention weight to enhance sparsity and constrain the attention mechanism to avoid redundant activation. The calculation method is expressed as:

[0205] ;

[0206] Where, is the number of network layers of the residual attention deep neural network;

[0207] is the number of attention heads;

[0208] For the residual attention deep neural network Tier The weight parameters of the attention heads.

[0209] Step 160: Based on the total loss function and the adaptive learning rate, the iterative trainable parameters are updated; when the preset stopping condition is met, a trained residual attention deep neural network is obtained.

[0210] Based on the total loss function and adaptive learning rate, the trainable parameters of the update iteration can be:

[0211] Based on the total loss function and adaptive learning rate, the network parameters are updated using the gradient clipping method. First, the gradient of the trainable parameters of the current round is calculated, and the gradient is clipped with the L2 norm to constrain its maximum amplitude not to exceed the preset threshold. Then, the momentum term is updated using the clipped gradient, and a weight decay term is added to suppress overfitting. Then, the trainable parameters of the next round of iteration are generated based on the updated momentum term. The specific steps are as follows:

[0212] First, the gradient of the trainable parameters is calculated as:

[0213] ;

[0214] Where, is the symbol of partial derivative;

[0215] are the trainable parameters of the residual attention deep neural network in the e-th iteration.

[0216] Furthermore, gradient clipping is used to constrain the gradient norm to prevent gradient explosion during training, which is expressed as:

[0217] ;

[0218] , is the clipping threshold, for example, set it to 5.0;

[0219] is the gradient of the trainable parameters after clipping;

[0220] is the L2 norm.

[0221] Furthermore, the momentum term is updated based on the gradient of the pruned trainable parameters, which is expressed as:

[0222] ;

[0223] Where, For the The momentum term of the round iteration is used to smooth the parameter update of the current round;

[0224] For the The momentum term of the round iteration;

[0225] is the momentum coefficient, for example, set to 0.9;

[0226] is the weight decay strength, e.g., set to .

[0227] Furthermore, the trainable parameters of the residual attention deep neural network are updated based on the momentum term, which is expressed as:

[0228] ;

[0229] Where, For the Trainable parameters of the iterative residual attention deep neural network.

[0230] When the preset stop iteration condition is met, the trained residual attention deep neural network is obtained, which can be specifically:

[0231] During the training process, the loss change on the validation set is continuously monitored. If the current training round reaches the preset maximum training round limit, the iteration is stopped immediately.

[0232] At the same time, an early stopping mechanism is used for dynamic judgment. That is, when the value of the total loss function does not show a significant decrease within the consecutive set rounds, which is manifested as not falling below the historical optimal verification loss by more than the preset threshold, or the adaptive learning rate has decayed to the preset minimum learning rate lower limit, the training process is also terminated.

[0233] Step 170: Obtain current monitoring data, calculate the enhanced features of the current monitoring data, and then use the trained residual attention deep neural network to evaluate the health status of the energy storage system.

[0234] Specifically, after the residual attention deep neural network training is completed, real-time health status classification is performed according to the following process:

[0235] 1) For the input sensor data stream (Current monitoring data), perform time-frequency dual-domain normalization, and output ;

[0236] 2) Calculate the fluctuation entropy vector ,and Splicing ;

[0237] 3) Input the residual attention deep neural network, pass through the residual attention module and the hybrid regularization layer in turn, and output the features after hybrid regularization processing ;

[0238] 4) Calculate the health state probability vector through the health state probabilistic output layer , take the index with the maximum probability as the predicted state.

[0239] In one embodiment, Figure 5 As shown, in order to verify the performance advantages of the residual attention deep neural network over the conventional model, the comparison objects include support vector machines, random forests, long short-term memory networks and convolutional neural networks. The bar chart clearly shows that the traditional model has a significantly low accuracy in the "mild degradation" and "moderate degradation" states, indicating that it is difficult to capture early weak degradation features. The conventional deep learning model has improved in the middle and late stages, but there are still omissions in the early degradation. The method of this application maintains high accuracy in all five health states, especially in the early degradation stage. The experimental results show that the residual attention module strengthens the degradation-sensitive features through dynamic adjustment of feature weights, the Markov output layer constrains the continuity of state transfer, and the dynamic focus loss function balances the sample imbalance problem, which together solves the defect of insufficient modeling of progressive degradation by conventional models.

[0240] In addition, this application Figure 6 This embodiment of the present application provides an energy storage system health status assessment system. Figure 6 As shown, the system provided in the embodiment of the present application mainly includes:

[0241] The acquisition module 210 is used to collect the original monitoring data of the energy storage system in real time and store the original monitoring data according to a preset time window;

[0242] Normalization calculation module 220 is used to perform time scale uniform processing on the original monitoring data within a preset time window to obtain a time normalization result, perform frequency domain processing on the original monitoring data to obtain a frequency domain compression result; and calculate the normalized data after time domain and frequency domain processing based on the time normalization result and the frequency domain compression result;

[0243] Enhanced calculation module 230, used to model the fluctuation characteristics of the time series based on the normalized data after time domain and frequency domain processing in combination with the fluctuation entropy mechanism, and calculate the characteristics of the enhanced monitoring data;

[0244] The learning rate adjustment module 240 is used to construct a residual attention deep neural network composed of an input layer, a plurality of residual attention modules, a hybrid regularization layer, and a state output layer stacked in sequence; by monitoring the variance change of the parameter gradient in each round of residual attention deep neural network training, the current adaptive learning rate is autonomously adjusted;

[0245] The loss calculation module 250 is used to output the health state probability of each health state category through the residual attention deep neural network; calculate the dynamic focus loss based on the health state probability of each health state category and the preset loss weight corresponding to each category; obtain the dynamic focus loss, temporal continuity loss, and attention sparse regularization loss to calculate the total loss function;

[0246] Iterative training module 260 is used to update iterative trainable parameters based on the total loss function and the adaptive learning rate; when a preset stopping condition is met, a trained residual attention deep neural network is obtained;

[0247] The real-time evaluation module 270 is used to obtain the current monitoring data, calculate the features of the enhanced current monitoring data, and then use the trained residual attention deep neural network to evaluate the health status of the energy storage system.

[0248] Thus far, the technical solutions of the present disclosure have been described in conjunction with the foregoing multiple embodiments. However, it is easy for those skilled in the art to understand that the scope of protection of the present disclosure is not limited to these specific embodiments. Without departing from the technical principles of the present disclosure, those skilled in the art may split and combine the technical solutions in the above-mentioned various embodiments, and may also make equivalent changes or replacements to the relevant technical features. Any changes, equivalent replacements, improvements, etc. made within the technical concepts and / or technical principles of the present disclosure will fall within the scope of protection of the present disclosure.

Claims

1. A method for evaluating the health status of an energy storage system, characterized in that: The method comprises: Collect the original monitoring data of the energy storage system in real time and store the original monitoring data according to the preset time window; The original monitoring data within the preset time window is processed in a unified time scale to obtain a time normalization result, and the original monitoring data is processed in the frequency domain to obtain a frequency domain compression result; the normalized data after time domain and frequency domain processing is calculated based on the time normalization result and the frequency domain compression result; Based on the normalized data after time domain and frequency domain processing, the fluctuation characteristics of the time series are modeled in combination with the fluctuation entropy mechanism to calculate the characteristics of the enhanced monitoring data; Construct a residual attention deep neural network consisting of an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer stacked in sequence; by monitoring the variance changes of the parameter gradients in each round of residual attention deep neural network training, the current adaptive learning rate is automatically adjusted; Through the residual attention deep neural network, the health status probability of each health status category is output; based on the health status probability of each health status category and the preset loss weight corresponding to each category, the dynamic focus loss is calculated; the dynamic focus loss, temporal continuity loss, and attention sparse regularization loss are obtained to calculate the total loss function; Based on the total loss function and adaptive learning rate, the iterative trainable parameters are updated; when the preset stopping conditions are met, the trained residual attention deep neural network is obtained; The current monitoring data is obtained, the features of the enhanced current monitoring data are calculated, and then the trained residual attention deep neural network is used to evaluate the health status of the energy storage system.

2. The method for evaluating the health status of an energy storage system according to claim 1, wherein: Real-time collection of raw monitoring data from the energy storage system, including: Real-time collection of raw monitoring data through a sensor network deployed on key components of the energy storage system; The data sources of the original monitoring data include the voltage sensor, current sensor, temperature sensor of the battery cell / module, and the charge and discharge status information recorded by the battery management system; The original monitoring data is collected using synchronous high-frequency sampling above 1 kHz.

3. The method for evaluating the health status of an energy storage system according to claim 1, wherein: The original monitoring data within the preset time window is processed in a unified time scale to obtain a time normalization result, and the original monitoring data is processed in the frequency domain to obtain a frequency domain compression result; The normalized data after time domain and frequency domain processing are calculated based on the time normalization results and frequency domain compression results, including: In the time dimension, the mean and standard deviation of the original monitoring data within the preset time window are calculated, and based on this, the data at the current moment is centered and scaled to unify the time scale and obtain the time normalization result; The time normalization result is mapped to the frequency domain space through the frequency domain transformation method to obtain the spectrum amplitude; the spectrum amplitude is compressed to the preset amplitude dynamic range using the logarithmic function; The time normalization result is multiplied element by element by the frequency domain compression result to obtain the normalized data after time domain and frequency domain processing.

4. The method for evaluating the health status of an energy storage system according to claim 1, wherein: Based on the normalized data after time domain and frequency domain processing, the fluctuation characteristics of the time series are modeled in combination with the fluctuation entropy mechanism, and the enhanced monitoring data features are calculated, including: Extracting normalized data at different time points through a sliding time window and calculating the distance between the extracted data and several pre-generated center vectors; According to the distance, the exponential mapping is used to construct the probability that the normalized data belongs to each center, and the fluctuation entropy vector corresponding to the normalized data is calculated based on the probability of each center; The normalized data is concatenated with the corresponding fluctuation entropy vector to obtain the enhanced monitoring data features.

5. The method for evaluating the health status of an energy storage system according to claim 1, wherein: Construct a residual attention deep neural network consisting of an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer stacked in sequence, specifically including: The input layer receives the features of the enhanced monitoring data and maps the features of the enhanced monitoring data to the initial high-dimensional feature space of the network; The residual path in the residual attention module extracts preset common features from the features of the enhanced monitoring data in the high-dimensional feature space, and the attention path dynamically adjusts the feature weights; through the preset common features, the features of the enhanced monitoring data, the self-attention weights, and the Sigmoid activation function, the features processed by the residual attention are obtained; The hybrid regularization layer randomly discards a preset part of the residual attention module output with a preset probability to obtain the first feature, and adopts the spectral normalization mechanism to perform spectral normalization on the residual attention module output to obtain the second feature; the feature after the hybrid regularization of the first feature and the second feature is obtained; The state output layer obtains the basic probability of each health state category through the conventional classification function; Based on the number differences between each health state category, a state transition matrix is constructed; The health state probability is calculated based on the basic probability, the features after hybrid regularization, and the state transition matrix.

6. The method for evaluating the health status of an energy storage system according to claim 1, wherein: By monitoring the variance changes of parameter gradients in each round of residual attention deep neural network training, the current adaptive learning rate is adjusted autonomously, including: Calculate the ratio of the gradient variance of the current round to the gradient variance of the previous round; where the current round is greater than 1; If the current gradient variance is large, the adaptive learning rate is relaxed; otherwise, the adaptive learning rate is gradually reduced; the relaxation / reduction of the learning rate is achieved through an exponential decay mechanism.

7. The method for evaluating the health status of an energy storage system according to claim 1, wherein: Based on the health status probability of each health status category and the preset loss weight corresponding to each category, the dynamic focus loss is calculated, specifically including: By formula: , calculate the dynamic focusing parameters of category c ;in, represents the number of samples in the cth category, c∈[1,C], C represents the total number of health status categories, represents the median number of samples in each category; By formula: , calculate the loss weight of category c ;in, represents the total number of samples, Indicates a preset constant; By formula: , calculate the dynamic focus loss ;in, represents the health status probability of category c.

8. The method for evaluating the health status of an energy storage system according to claim 1, wherein: Obtain dynamic focus loss, temporal continuity loss, attention sparse regularization loss, and calculate the total loss function, including: Obtain the corresponding weight values of dynamic focus loss, temporal continuity loss, and attention sparse regularization loss, and then weight them to obtain the total loss function.

9. The method for evaluating the health status of an energy storage system according to claim 1, wherein: Based on the total loss function and adaptive learning rate, the trainable parameters of the update iteration are as follows: Calculate the gradient of the trainable parameters using the total loss function and the trainable parameters of the current round; Adopt gradient clipping to clip the gradient of trainable parameters; Use the clipped gradient and adaptive learning rate to update the momentum term, and add a preset weight decay term to suppress overfitting; Generate the trainable parameters for the next iteration based on the current momentum term, the updated momentum term, the adaptive learning rate, the current round of trainable parameters, and the gradient of the pruned trainable parameters.

10. A system for evaluating the health status of an energy storage system, characterized in that: The system comprises: The acquisition module is used to collect the original monitoring data of the energy storage system in real time and store the original monitoring data according to the preset time window; The normalization calculation module is used to perform time-scale uniform processing on the original monitoring data within the preset time window to obtain a time normalization result, and perform frequency domain processing on the original monitoring data to obtain a frequency domain compression result; the normalized data after time domain and frequency domain processing is calculated based on the time normalization result and the frequency domain compression result; The enhanced computing module is used to model the fluctuation characteristics of the time series based on the normalized data processed in the time and frequency domains, combined with the fluctuation entropy mechanism, and calculate the characteristics of the enhanced monitoring data; The learning rate adjustment module is used to construct a residual attention deep neural network composed of an input layer, several residual attention modules, a hybrid regularization layer, and a state output layer. By monitoring the variance changes of the parameter gradients in each round of residual attention deep neural network training, the current adaptive learning rate is automatically adjusted. The loss calculation module is used to output the health status probability of each health status category through the residual attention deep neural network; calculate the dynamic focus loss based on the health status probability of each health status category and the preset loss weight corresponding to each category; obtain the dynamic focus loss, temporal continuity loss, and attention sparse regularization loss to calculate the total loss function; The iterative training module is used to update the iterative trainable parameters based on the total loss function and adaptive learning rate; when the preset stopping condition is met, the trained residual attention deep neural network is obtained; The real-time evaluation module is used to obtain current monitoring data, calculate the enhanced features of the current monitoring data, and then use the trained residual attention deep neural network to evaluate the health status of the energy storage system.

Images