Hybrid energy storage battery state monitoring method and system based on big data

Abstract

The invention discloses a hybrid energy storage battery state monitoring method and system based on big data, and relates to the technical field of battery state monitoring, and the method comprises the steps: collecting battery state data, and carrying out the preprocessing of the battery state data; based on the preprocessed battery state data, a high-dimensional feature vector is obtained through dynamic weight adjustment and fusion; constructing a battery state prediction model based on the long and short term memory network model; obtaining a battery health value through a battery state prediction model based on the high-dimensional feature vector; and monitoring the health state of the battery in real time based on the health value of the battery. According to the method, the weight is dynamically adjusted and the high-dimensional feature vector is generated through fusion, so that the prediction precision of the health value of the battery is remarkably improved, the real-time monitoring and fault early warning of the health state of the battery under the complex working condition are realized, and the safety, the reliability and the operation management efficiency of the hybrid energy storage system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of battery status monitoring, and in particular to a hybrid energy storage battery status monitoring method and system based on big data. Background Art

[0002] With the development of energy storage technology, batteries, as the core components of energy storage systems, have been widely used in power systems, new energy vehicles, smart grids and other fields. Especially in hybrid energy storage systems, batteries work together with other energy storage devices (such as supercapacitors) to provide more stable and efficient energy management for complex operating conditions. However, the performance of batteries will gradually deteriorate with the extension of usage time, and their operating status is affected by many factors, including electrical properties (such as voltage, current, internal resistance, etc.), environmental conditions (such as temperature, humidity, etc.), and load changes. Therefore, real-time monitoring of the operating status of batteries and predicting their health status have become important technical means to ensure the safety and reliability of energy storage systems. In recent years, battery status monitoring methods based on big data have gradually become a research hotspot. The core lies in using large-scale operating data to comprehensively analyze the operating status of batteries and achieve health management through technologies such as feature extraction, data modeling and status prediction.

[0003] In the prior art, there are certain limitations in the monitoring methods for battery status. Traditional methods often rely on a single feature (such as voltage or temperature) for simple threshold judgments, which makes it difficult to fully reflect the operating status of the battery. In addition, although some methods have introduced multidimensional data analysis, the weight distribution between different features is unreasonable, resulting in the failure to fully reflect the role of key influencing factors, reducing the accuracy of monitoring. At the same time, when predicting the health status of batteries, the existing methods lack in-depth mining of the dynamic change laws in time series data, and cannot effectively capture the short-term and long-term status change trends of the battery. These deficiencies make the existing battery status monitoring technology show certain lags and inaccuracies when dealing with complex operating conditions. Summary of the invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a hybrid energy storage battery status monitoring method based on big data to solve the problems of insufficient accuracy of feature extraction and fusion and low accuracy of health status prediction.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides a hybrid energy storage battery status monitoring method based on big data, which comprises:

[0008] Collect battery status data and preprocess the battery status data; based on the preprocessed battery status data, obtain high-dimensional feature vectors by dynamically adjusting weights and fusion; build a battery status prediction model based on the long short-term memory network model; obtain the battery health value through the battery status prediction model based on the high-dimensional feature vector; monitor the battery health status in real time based on the battery health value.

[0009] As a preferred solution of the hybrid energy storage battery status monitoring method based on big data described in the present invention, wherein: the battery status data includes electrical characteristic data, environmental parameters and operating load information;

[0010] The preprocessing of the battery status data includes removing outliers, interpolating missing values ​​and normalizing.

[0011] As a preferred solution of the hybrid energy storage battery state monitoring method based on big data of the present invention, wherein: based on the pre-processed battery state data, high-dimensional feature vectors are obtained by dynamically adjusting weights and fusion, and the specific steps are as follows:

[0012] Extracting battery status features from the preprocessed battery status data by using statistical analysis methods;

[0013] Through the correlation analysis method, the influence of the pre-processed battery status data on the battery performance is analyzed, and the influencing factors affecting the battery performance are extracted;

[0014] Evaluate the pre-processed battery status data through statistical methods to obtain credibility;

[0015] Based on the influencing factors and credibility, the battery status feature weight is dynamically adjusted, and the expression is:

[0016] Wi(s)=γ1×Ei(s)+γ2×Ri(s);

[0017] Wherein, Wi(s) is the weight of the i-th battery state feature at time s after dynamic adjustment, Ei(s) represents the influencing factor of the i-th battery state feature at time s, Ri(s) represents the credibility of the i-th battery state feature at time s, γ1 is the weight coefficient of Ei(s), γ2 is the weight coefficient of Ri(s), s is the time index variable, and i is the battery state feature index variable;

[0018] Through physical constraints, the weight Wi(s) of the dynamically adjusted i-th battery state feature at time s is fused to obtain a high-dimensional feature vector.

[0019] As a preferred solution of the hybrid energy storage battery state monitoring method based on big data described in the present invention, wherein: through physical constraints, the weight Wi(s) of the dynamically adjusted i-th battery state feature at time s is fused to obtain a high-dimensional feature vector. The specific steps are as follows:

[0020] Define physical constraint rules through battery operation mechanism, operating conditions, safety thresholds and experimental data;

[0021] Define correction factors based on physical constraint rules;

[0022] Correct the dynamically adjusted weights through correction factors;

[0023] Use the sliding window method to fuse the corrected weights of the current time point and the historical time point to obtain a high-dimensional feature vector, expressed as:

[0024]

[0025] Among them, V(s) is a high-dimensional feature vector, Q i (sk) is the weight of the i-th battery state feature after correction at time sk, X i (sk) is the battery status feature of the ith battery at time sk, k is the index variable of the time lag, and N is the total number of included time points.

[0026] As a preferred solution of the hybrid energy storage battery state monitoring method based on big data described in the present invention, wherein: the battery state prediction model is constructed based on the long short-term memory network model, and the specific steps are as follows:

[0027] The model is based on the long short-term memory network model;

[0028] Arrange the high-dimensional historical feature vectors in chronological order and use the sliding window method to create the input time series;

[0029] The input layer of the LSTM model accepts the input time series;

[0030] The LSTM multi-layer structure is used to capture the dynamic change relationship in the time series and extract the time-dependent features;

[0031] The fully connected layer maps the output time-dependent features to the target feature dimension;

[0032] Build a battery status prediction model.

[0033] As a preferred solution of the hybrid energy storage battery state monitoring method based on big data of the present invention, wherein: the battery health value is obtained through the battery state prediction model based on the high-dimensional feature vector, and the specific steps are as follows:

[0034] Organize high-dimensional feature vectors into a sliding window format in chronological order to form a time series input;

[0035] The multi-layer structure captures the dynamic change relationship between short-term and long-term. The first layer of LSTM extracts the low-level dynamic characteristics of the time series, and the second layer of LSTM further extracts the time-dependent features.

[0036] The output time-dependent features of LSTM are passed to the fully connected layer to obtain the battery health value, expressed as:

[0037]

[0038] Where Y(s) is the battery health value at time s, σ is the activation function, H is the total number of neurons in the fully connected layer, and w j is the weight of the neuron in the jth fully connected layer, F is the dimension of the high-dimensional feature vector, Q i (s) is the weight of the i-th battery state feature after correction at time s, X i (s) is the battery status feature of the ith battery at time s, α is a parameter, b j is the bias term of the jth fully connected layer neuron, c is a constant term, i is the battery state feature index variable, and j is the index variable of the fully connected layer neuron.

[0039] As a preferred solution of the hybrid energy storage battery status monitoring method based on big data described in the present invention, wherein: the battery health status is monitored in real time based on the battery health value, and the specific steps are as follows:

[0040] Based on the average value of the battery health value, a threshold T is set;

[0041] Compare the battery health value Y(s) with the threshold T to monitor the battery health status in real time;

[0042] When Y(s)>T, the battery is in good health and continues to operate normally;

[0043] When Y(s)≤T, the battery health status is poor and corresponding measures should be taken immediately.

[0044] In the second aspect, the present invention provides a hybrid energy storage battery status monitoring system based on big data, including a data acquisition and preprocessing module, a feature generation module, a model construction module, a health value calculation module and a status monitoring module; the data acquisition and preprocessing module is used to collect battery status data and preprocess the battery status data; the feature generation module is used to obtain a high-dimensional feature vector based on the preprocessed battery status data by dynamically adjusting weights and fusion; the model construction module is used to construct a battery status prediction model based on a long short-term memory network model; the health value calculation module is used to obtain a battery health value through a battery status prediction model based on a high-dimensional feature vector; the status monitoring module is used to monitor the battery health status in real time based on the battery health value.

[0045] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the hybrid energy storage battery status monitoring method based on big data as described in the first aspect of the present invention is implemented.

[0046] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the hybrid energy storage battery status monitoring method based on big data as described in the first aspect of the present invention is implemented.

[0047] The beneficial effects of the present invention are as follows: the present invention solves the problem of inaccurate multi-dimensional feature analysis and fusion by dynamically adjusting weights and fusing to generate high-dimensional feature vectors, and constructs a battery status prediction model based on a long short-term memory network (LSTM), deeply explores the dynamic change rules in time series, and significantly improves the accuracy of battery health value prediction, realizes real-time monitoring of battery health status and fault warning under complex working conditions, and effectively improves the safety, reliability and operation management efficiency of the hybrid energy storage system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0049] Figure 1 This is a flow chart of the hybrid energy storage battery status monitoring method based on big data in Example 1.

[0050] Figure 2 This is a system diagram of the hybrid energy storage battery status monitoring system based on big data in Example 1. DETAILED DESCRIPTION

[0051] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0052] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0053] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0054] Example 1, reference Figure 1 and Figure 2 , which is the first embodiment of the present invention, and provides a hybrid energy storage battery status monitoring method based on big data, comprising the following steps:

[0055] S1. The battery status data includes electrical characteristic data, environmental parameters and operating load information.

[0056] S1.1. The preprocessing of the battery status data includes removing outliers, interpolating missing values ​​and normalizing.

[0057] It should be noted that the removal of outliers is to identify abnormal data points (such as sudden changes in voltage, current or temperature values) in the battery status data that deviate from the normal range through rule detection, and remove or replace them;

[0058] Missing value interpolation is to fill in the missing values ​​caused by equipment failure or transmission problems during the collection process through interpolation methods (such as mean interpolation, linear interpolation or time series prediction methods) to restore the integrity of the data;

[0059] Normalization is the process of normalizing different physical quantities (such as voltage, current, temperature, etc.) to a unified numerical range (such as [0, 1] or [-1, 1]) to eliminate the impact of feature dimension differences on model training.

[0060] S2. Based on the preprocessed battery status data, a high-dimensional feature vector is obtained by dynamically adjusting weights and fusion.

[0061] S2.1. Extract battery status features from the preprocessed battery status data through statistical analysis methods.

[0062] It should be noted that the battery status characteristics include central tendency characteristics, dispersion degree characteristics and distribution morphology characteristics;

[0063] It should also be noted that, based on the preprocessed battery status data, the mean is obtained by summing up all data points and then dividing the sum by the number of data points;

[0064] Sort the data points by numerical value. If the number of data points is an odd number, the median is the data point in the middle; if the number of data points is an even number, the median is the average of the two middle data points.

[0065] Through the mean and median, we can obtain the central tendency characteristics that reflect the central position of the data set;

[0066] Calculate the square of the difference between each data point and the mean, and then average these squared differences to get the variance;

[0067] The positive square root of the variance is the standard deviation;

[0068] The range is obtained by subtracting the minimum value from the maximum value in the data;

[0069] Through variance, standard deviation and range, the discrete degree characteristics reflecting the degree of data dispersion are obtained;

[0070] The skewness is obtained by calculating the difference between each data point and the mean and normalizing it (dividing it by the standard deviation). The normalized differences are then cubed and averaged to obtain the skewness.

[0071] The kurtosis is obtained by taking the fourth power of the standardized difference and taking the average, then subtracting 3;

[0072] Through skewness and kurtosis, the distribution morphology characteristics that reflect the shape of data distribution are obtained.

[0073] S2.2. Analyze the influence of the pre-processed battery status data on the battery performance through correlation analysis method, and extract the influencing factors affecting the battery performance.

[0074] It should be noted that the factors affecting battery performance are extracted as follows:

[0075] Calculate the Pearson correlation coefficient between the preprocessed battery status data and another preprocessed battery status data, the expression is:

[0076]

[0077] Among them, r xy is the Pearson correlation coefficient, x z is the value of the zth preprocessed battery status data, y zis the zth value of another preprocessed battery status data, is all x z The mean of the values, All y z The mean of n is the total number of observed data points (x z and z Combined into one observation data point), z is the index variable of the observation data point;

[0078] The calculated correlation coefficient r xy The value range is between [-1,1], evaluate x z and z The linear correlation between

[0079] When xy >0, it means x z and z Positive correlation;

[0080] When xy <0, it means x z and z Negative correlation;

[0081] When xy =0, indicating x z and z There is no obvious linear relationship;

[0082] Based on the Pearson correlation coefficient r xy , determine the p value by t-test method, and judge x z and z The significance of the correlation is expressed as:

[0083]

[0084] Where t is the value used to judge the significance of the correlation, n is the total number of observed data points, and n-2 is the degree of freedom;

[0085] The p-value is determined by the t distribution table. If the p-value is less than the significance level (usually 0.05), then x is considered z and z The correlation is significant;

[0086] It should be noted that the p-value is a probability value for judging the significance of the hypothesis test results;

[0087] The significantly correlated variables were taken as the influencing factors affecting the battery performance.

[0088] S2.3. Evaluate the pre-processed battery status data through statistical methods to obtain credibility.

[0089] It should be noted that the credibility is obtained as follows:

[0090] Based on the Pearson correlation coefficient r xy , calculate the 95% confidence interval of the correlation coefficient to assess the stability of the correlation, the expression is:

[0091]

[0092] Where CI is the confidence interval, t α / 2,n-2 is the critical value of the t distribution, α is the significance level, and n-2 is the degree of freedom;

[0093] When x z and z When the correlation is significant and the confidence interval CI is narrow, the variable correlation is considered to have a high degree of credibility.

[0094] S2.4. Based on the influencing factors and credibility, the battery status feature weight is dynamically adjusted, and the expression is:

[0095] Wi(s)=γ1×Ei(s)+γ2×Ri(s);

[0096] Among them, Wi(s) is the dynamically adjusted weight of the i-th battery state feature at time s, Ei(s) represents the influencing factor of the i-th battery state feature at time s, Ri(s) represents the credibility of the i-th battery state feature at time s, γ1 is the weight coefficient of Ei(s), γ2 is the weight coefficient of Ri(s), s is the time index variable, and i is the battery state feature index variable.

[0097] S3. Through physical constraints, the weight Wi(s) of the dynamically adjusted i-th battery state feature at time s is fused to obtain a high-dimensional feature vector.

[0098] S3.1. Define physical constraint rules based on battery operation mechanism, operating conditions, safety thresholds and experimental data.

[0099] It should be noted that the battery operating mechanism, operating conditions, safety thresholds and experimental data are derived from the battery manufacturer's technical specifications, industry standards, actual operating data and laboratory test results.

[0100] It should also be noted that the physical constraint rules are as follows:

[0101] Temperature: In high-rate discharge or extreme environments, the importance (weight) of temperature should be increased;

[0102] Voltage: When the battery is fast charged, the importance of voltage should be enhanced;

[0103] Current: During steady-state operation, the current weighting may be reduced.

[0104] Specifically, through the battery operation mechanism, the inherent physical relationship between battery status data (such as voltage, current, temperature, etc.) and its impact on battery performance are analyzed; combined with operating conditions, the importance of each battery status data under different operating conditions (such as high-rate discharge, fast charging, and static) is evaluated; at the same time, based on the safety threshold, the physical limit range of key parameters (such as temperature, voltage, current) is set to avoid abnormal operation; finally, through experimental data verification and induction, the weight adjustment rules of each battery status data under different conditions are summarized to form a complete physical constraint rule.

[0105] S3.2. Define correction factors based on physical constraint rules.

[0106] It should be noted that these physical constraints are analyzed through statistical analysis methods, the key factors affecting the battery state characteristics are identified, and the correction factors are set accordingly.

[0107] S3.3. Correct the dynamically adjusted weights using correction factors.

[0108] It should be noted that the dynamically adjusted weight is corrected by the correction factor, and the expression is:

[0109] Q i (s) = W i (s) × L;

[0110] Among them, Q i (s) is the modified weight and L is the modification factor.

[0111] S3.4. Use the sliding window method to fuse the corrected weights of the current time point and the historical time point to obtain a high-dimensional feature vector, expressed as:

[0112]

[0113] Among them, V(s) is a high-dimensional feature vector, Q i (sk) is the weight of the i-th battery state feature after correction at time sk, X i (sk) is the battery status feature of the ith battery at time sk, k is the index variable of the time lag, and N is the total number of included time points.

[0114] S4. Build a battery status prediction model based on the long short-term memory network model.

[0115] S4.1, based on the long short-term memory network model;

[0116] Arrange the high-dimensional historical feature vectors in chronological order and use the sliding window method to create the input time series;

[0117] The input layer of the LSTM model accepts the input time series;

[0118] The LSTM multi-layer structure is used to capture the dynamic change relationship in the time series and extract the time-dependent features;

[0119] The fully connected layer maps the output time-dependent features to the target feature dimension;

[0120] Build a battery status prediction model.

[0121] It should be noted that the advantage of the long short-term memory network (LSTM) model as the basic model is that it can effectively capture the long-term dependencies and dynamic changes in time series data, process high-dimensional historical feature vectors through a multi-layer structure, and extract time-dependent features. The LSTM model is particularly suitable for processing data with time sequence, and can remember past information and use it for future predictions, thereby improving the accuracy and robustness of battery status predictions.

[0122] S5. Obtain the battery health value through the battery status prediction model based on the high-dimensional feature vector.

[0123] S5.1. Organize the high-dimensional feature vector into a sliding window format in time order to form a time series input.

[0124] S5.2. The short-term and long-term dynamic change relationships are captured through a multi-layer structure. The first layer of LSTM extracts the low-level dynamic characteristics of the time series, and the second layer of LSTM further extracts the time-dependent features.

[0125] S5.3. The output time-dependent features of LSTM are passed to the fully connected layer to obtain the battery health value, which is expressed as:

[0126]

[0127] Where Y(s) is the battery health value at time s, σ is the activation function, H is the total number of neurons in the fully connected layer, and w j is the weight of the neuron in the jth fully connected layer, F is the dimension of the high-dimensional feature vector, Q i (s) is the weight of the i-th battery state feature after correction at time s, X i (s) is the battery status feature of the ith battery at time s, α is a parameter, b j is the bias term of the jth fully connected layer neuron, c is a constant term, i is the battery state feature index variable, and j is the index variable of the fully connected layer neuron.

[0128] It should be noted that the calculation process of the expression is as follows:

[0129] For each input feature Qi (s) and X i (s), calculate its value after nonlinear transformation, sum these transformed values, and add the bias term b j , and then activated by the hyperbolic tangent function, multiplying each activated result by the corresponding weight w j , and sum the results of all neurons, transform the summed result nonlinearly through an activation function, and add the constant term c to get the final output Y(s).

[0130] S6. Based on the battery health value, monitor the battery health status in real time.

[0131] S6.1. Set a threshold T based on the average value of the battery health value.

[0132] S6.2, compare the battery health value Y(s) with the threshold T, and monitor the battery health status in real time;

[0133] When Y(s)>T, the battery is in good health and continues to operate normally;

[0134] When Y(s)≤T, the battery health status is poor and corresponding measures should be taken immediately.

[0135] It should be noted that the corresponding measures include reducing the load, adjusting the operating conditions and replacing the battery.

[0136] This embodiment also provides a hybrid energy storage battery status monitoring system based on big data, including: a data acquisition and preprocessing module, a feature generation module, a model building module, a health value calculation module and a status monitoring module;

[0137] A data acquisition and preprocessing module is used to acquire battery status data and preprocess the battery status data;

[0138] A feature generation module is used to obtain a high-dimensional feature vector based on the preprocessed battery status data by dynamically adjusting weights and fusing;

[0139] A model building module is used to build a battery state prediction model based on a long short-term memory network model;

[0140] A health value calculation module is used to obtain the battery health value through a battery state prediction model based on a high-dimensional feature vector;

[0141] The status monitoring module is used to monitor the battery health status in real time based on the battery health value.

[0142] This embodiment also provides a computer device, which is suitable for the hybrid energy storage battery status monitoring method based on big data, including: a memory and a processor; the memory is used to store computer executable instructions, and the processor is used to execute computer executable instructions to implement the hybrid energy storage battery status monitoring method based on big data proposed in the above embodiment.

[0143] The computer device may be a terminal, and the computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a key, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse, etc.

[0144] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the hybrid energy storage battery state monitoring method based on big data proposed in the above embodiment is implemented; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, referred to as EPROM), programmable read-only memory (Programmable Red-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic storage, flash memory, disk or optical disk.

[0145] In summary, the present invention solves the problem of inaccurate multi-dimensional feature analysis and fusion by dynamically adjusting weights and fusing to generate high-dimensional feature vectors, and constructs a battery status prediction model based on a long short-term memory network (LSTM), deeply explores the dynamic change rules in time series, and significantly improves the accuracy of battery health value prediction, realizes real-time monitoring of battery health status and fault warning under complex working conditions, and effectively improves the safety, reliability and operation management efficiency of the hybrid energy storage system.

[0146] 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 preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A hybrid energy storage battery status monitoring method based on big data, characterized in that: include, Collect battery status data and pre-process the battery status data; Based on the preprocessed battery status data, high-dimensional feature vectors are obtained by dynamically adjusting weights and fusion; Based on the long short-term memory network model, a battery status prediction model is constructed; Obtain the battery health value through the battery status prediction model based on the high-dimensional feature vector; Based on the battery health value, monitor the battery health status in real time.

2. The hybrid energy storage battery status monitoring method based on big data according to claim 1, characterized in that: The battery status data includes electrical characteristic data, environmental parameters and operating load information; The preprocessing of the battery status data includes removing outliers, interpolating missing values ​​and normalizing.

3. The hybrid energy storage battery status monitoring method based on big data as claimed in claim 2, characterized in that: Based on the pre-processed battery status data, high-dimensional feature vectors are obtained by dynamically adjusting weights and fusion. The specific steps are as follows: Extracting battery status features from the preprocessed battery status data by using statistical analysis methods; Through the correlation analysis method, the influence of the pre-processed battery status data on the battery performance is analyzed, and the influencing factors affecting the battery performance are extracted; Evaluate the pre-processed battery status data through statistical methods to obtain credibility; Based on the influencing factors and credibility, the battery status feature weight is dynamically adjusted, and the expression is: Wi(s)=γ1×Ei(s)+γ2×Ri(s); Wherein, Wi(s) is the weight of the i-th battery state feature at time s after dynamic adjustment, Ei(s) represents the influencing factor of the i-th battery state feature at time s, Ri(s) represents the credibility of the i-th battery state feature at time s, γ1 is the weight coefficient of Ei(s), γ2 is the weight coefficient of Ri(s), s is the time index variable, and i is the battery state feature index variable; Through physical constraints, the weight Wi(s) of the dynamically adjusted i-th battery state feature at time s is fused to obtain a high-dimensional feature vector.

4. The hybrid energy storage battery status monitoring method based on big data as claimed in claim 3 is characterized in that: Through physical constraints, the weight Wi(s) of the dynamically adjusted i-th battery state feature at time s is fused to obtain a high-dimensional feature vector. The specific steps are as follows: Define physical constraint rules through battery operation mechanism, operating conditions, safety thresholds and experimental data; Define correction factors based on physical constraint rules; Correct the dynamically adjusted weights through correction factors; Use the sliding window method to fuse the corrected weights of the current time point and the historical time point to obtain a high-dimensional feature vector, expressed as: Among them, V(s) is a high-dimensional feature vector, Q i (sk) is the weight of the i-th battery state feature after correction at time sk, X i (sk) is the battery status feature of the ith battery at time sk, k is the index variable of the time lag, and N is the total number of included time points.

5. The hybrid energy storage battery status monitoring method based on big data according to claim 4, characterized in that: The battery status prediction model is constructed based on the long short-term memory network model. The specific steps are as follows: The model is based on the long short-term memory network model; Arrange the high-dimensional historical feature vectors in chronological order and use the sliding window method to create the input time series; The input layer of the LSTM model accepts the input time series; The LSTM multi-layer structure is used to capture the dynamic change relationship in the time series and extract the time-dependent features; The fully connected layer maps the output time-dependent features to the target feature dimension; Build a battery status prediction model.

6. The hybrid energy storage battery status monitoring method based on big data according to claim 5, characterized in that: The specific steps of obtaining the battery health value through the battery status prediction model based on the high-dimensional feature vector are as follows: The multi-layer structure captures the dynamic change relationship between short-term and long-term. The first layer of LSTM extracts the low-level dynamic characteristics of the time series, and the second layer of LSTM further extracts the time-dependent features. The output time-dependent features of LSTM are passed to the fully connected layer to obtain the battery health value, expressed as: Where Y(s) is the battery health value at time s, σ is the activation function, H is the total number of neurons in the fully connected layer, and w j is the weight of the neuron in the jth fully connected layer, F is the dimension of the high-dimensional feature vector, Q i (s) is the weight of the i-th battery state feature after correction at time s, X i (s) is the battery status feature of the ith battery at time s, α is a parameter, b j is the bias term of the jth fully connected layer neuron, c is a constant term, i is the battery state feature index variable, and j is the index variable of the fully connected layer neuron.

7. The hybrid energy storage battery status monitoring method based on big data according to claim 6, characterized in that: The specific steps of real-time monitoring of the battery health status based on the battery health value are as follows: Based on the average value of the battery health value, a threshold T is set; Compare the battery health value Y(s) with the threshold T to monitor the battery health status in real time; When Y(s)>T, the battery is in good health and continues to operate normally; When Y(s)≤T, the battery health status is poor and corresponding measures should be taken immediately.

8. A hybrid energy storage battery status monitoring system based on big data, based on the hybrid energy storage battery status monitoring method based on big data according to any one of claims 1 to 7, characterized in that: Including data acquisition and preprocessing module, feature generation module, model building module, health value calculation module and status monitoring module; A data acquisition and preprocessing module is used to acquire battery status data and preprocess the battery status data; A feature generation module is used to obtain a high-dimensional feature vector based on the preprocessed battery status data by dynamically adjusting weights and fusing; A model building module is used to build a battery state prediction model based on a long short-term memory network model; A health value calculation module is used to obtain the battery health value through a battery state prediction model based on a high-dimensional feature vector; The status monitoring module is used to monitor the battery health status in real time based on the battery health value.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the hybrid energy storage battery status monitoring method based on big data described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the hybrid energy storage battery status monitoring method based on big data described in any one of claims 1 to 7 are implemented.

Images