A working state monitoring system and method for energy storage devices used in industry and commerce

Abstract

The application relates to a kind of working state monitoring system and method of commercial energy storage device, system includes: data acquisition module, for obtaining real-time acquisition commercial energy storage device temperature data, initial voltage data and initial current data, and correcting initial voltage data and initial current data;Data transmission module is used for packing compression temperature data, processed voltage data and current data, and the data transmission to state monitoring module after packing compression;SOC estimation module is used for inputting voltage data and current data into SOC estimation model, and obtaining SOC estimation result;SOH estimation module is used for inputting temperature data, voltage data and current data into SOH estimation model, and obtaining SOH estimation result;State monitoring module is used for obtaining the working state of commercial energy storage device according to SOC estimation result and SOH estimation result.The application can improve the accuracy of working state monitoring.

Description

Technical Field

[0001] This invention relates to the field of energy storage device monitoring technology, and in particular to a system and method for monitoring the operating status of industrial and commercial energy storage devices. Background Technology

[0002] The industrial and commercial sector constitutes my country's largest electricity market, characterized by high and volatile electricity prices. With the help of energy storage systems, industrial and commercial users can charge their devices during off-peak hours and discharge them during peak hours, avoiding periods of high electricity prices and thus reducing electricity costs. By optimizing electricity prices, cost savings and improved economic efficiency can be achieved. Installing energy storage systems not only reduces energy costs and improves energy utilization efficiency, but also improves power quality, enhances emergency backup capabilities, and achieves energy self-sufficiency, among other benefits. With the continuous development of energy storage technology and the growth of market demand, the application of energy storage systems in the industrial and commercial sectors will be further promoted and popularized.

[0003] Monitoring the operating status of commercial and industrial energy storage devices helps improve their safety, reliability, and economy. However, current monitoring of the operating status of commercial and industrial energy storage devices faces several challenges: Traditional energy storage data acquisition methods inherently contain errors in voltage and current measurements, and voltage varies to different degrees under different operating conditions due to changes in ambient temperature, resulting in highly variable data. Existing methods for estimating battery SOC, such as data-driven methods like neural networks and regression analysis, do not rely on high-precision battery models but require extensive data acquisition. Training results depend on sample quality and are prone to overfitting and poor generalization. Regarding SOH estimation, neural networks are widely used in power battery SOH estimation due to their efficient data processing and feature extraction capabilities. Model parameters significantly impact regression accuracy, but model accuracy is low and generalization ability is poor. Therefore, effectively acquiring and processing data and accurately identifying the operating status of energy storage devices are problems that need to be solved.

[0004] To address the above problems, the present invention provides a system and method for monitoring the operating status of industrial and commercial energy storage devices. Summary of the Invention

[0005] The purpose of this invention is to provide a system and method for monitoring the operating status of industrial and commercial energy storage devices, which can improve the accuracy of operating status monitoring.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] A working status monitoring system for an industrial and commercial energy storage device includes: a data acquisition module, a data transmission module, a SOC prediction module, a SOH prediction module, and a status monitoring module;

[0008] The data acquisition module is used to acquire monitoring data, wherein the monitoring data is real-time acquired temperature data, initial voltage data, and initial current data of industrial and commercial energy storage devices, and to perform correction processing on the initial voltage data and the initial current data.

[0009] The data transmission module is used to package and compress temperature data, processed voltage data and current data, and transmit the packaged and compressed data to the status monitoring module.

[0010] The SOC prediction module is used to input the voltage data and the current data into the SOC prediction model to obtain the SOC prediction result, wherein the SOC prediction model is constructed based on a neural network and a battery electrochemical model.

[0011] The SOH prediction module is used to input the temperature data, voltage data and current data into the SOH prediction model to obtain the SOH prediction result. The SOH prediction model is constructed based on support vector regression optimized by particle swarm optimization.

[0012] The status monitoring module is used to obtain the operating status of the industrial and commercial energy storage device based on the SOC prediction result and the SOH prediction result.

[0013] Optionally, the data acquisition module includes a calibration model construction unit and a data calibration unit;

[0014] The correction model construction unit is used to acquire historical voltage data and historical current data, and input the initial voltage correction model and the initial current correction model to calculate the voltage correction parameters and current correction parameters, and to obtain the voltage correction model and the current correction model.

[0015] The data correction unit is used to correct the initial voltage data and initial current data acquired in real time using the voltage correction model and the current correction model, respectively.

[0016] Optionally, obtaining historical voltage and current data includes:

[0017] Initial historical voltage data and initial historical current data are collected within a preset time period by a preset sampling rate. Missing values ​​are filled and outliers are replaced in the initial historical voltage data and initial historical current data to obtain the historical voltage data and the historical current data. For missing values ​​and outliers, the mean of the two adjacent data points of the missing value and the outlier are used for filling and replacement respectively.

[0018] Optionally, the method for correcting the real-time acquired initial voltage data using the voltage correction model is as follows:

[0019] U = [(U' / c) – b] × a

[0020] Where U is the corrected voltage value, U' is the voltage value acquired before correction, a is the voltage amplification factor, b is the first DC bias, c is the reference source factor, c = C' / C, where C is the reference source voltage and C' is the reference voltage read by the system;

[0021] The method for correcting the real-time acquired initial current data using the current correction model is as follows:

[0022] I = [(I' / c–d)] × e

[0023] Where I is the corrected current value, I' is the current value collected before correction, e is the current amplification factor, and d is the second DC bias.

[0024] Optionally, the SOC prediction module includes: a model fitting unit, a data processing unit, and an SOC prediction unit;

[0025] The model fitting unit is used to fit the battery electrochemical model using a neural network to obtain the discrete equation of the SOC prediction model.

[0026] The data processing unit is used to process the current data, the voltage data, and the discrete equations using an improved unscented Kalman filter algorithm to obtain the Kalman gain.

[0027] The SOC prediction unit is used to estimate the parameters of the industrial and commercial energy storage device at the current moment based on the Kalman gain, and obtain the SOC prediction result at the current moment.

[0028] Optionally, fitting the battery electrochemical model using a neural network includes: fitting the nonlinear discrete-time state-space model of the battery electrochemical model using a BP neural network; and fitting the SOC-Voc curve of the battery electrochemical model using a high-order neural network.

[0029] Optionally, the SOH prediction module includes an SOH prediction model construction unit;

[0030] The SOH prediction model construction unit is used to train the SVR model based on the PSO algorithm and update the individual extreme values ​​and global extreme values ​​of the particle swarm, determine the optimal values ​​of the parameters, and construct an SVR regression model based on the optimal values ​​as the SOH prediction model.

[0031] Optionally, training the SVR model based on the PSO algorithm and updating the individual and global extrema of the particle swarm includes:

[0032] S1. Use the mean squared error of the SVR regression model as the fitness function, and initialize the particle swarm position and velocity;

[0033] S2. Train the SVR regression model using sample data, calculate particle fitness, and obtain the globally optimal initial value based on the results;

[0034] S3. Update the position and velocity of the particles;

[0035] S4. Train the SVR regression model and update the individual extreme values ​​and global extreme values ​​of the particles;

[0036] S5. Determine if the termination condition has been met. If the number of iterations reaches the set threshold, the current global extreme value is used as the optimization output; otherwise, return to S3 to continue the iteration.

[0037] On the other hand, a method for monitoring the operating status of industrial and commercial energy storage devices is also provided, including:

[0038] Real-time acquisition of temperature data, initial voltage data, and initial current data of industrial and commercial energy storage devices, and correction processing of the initial voltage data and the initial current data;

[0039] The temperature data, processed voltage data, and current data are packaged and compressed, and the packaged and compressed data is transmitted to the SOC prediction model and the SOH prediction model.

[0040] The voltage data and the current data are input into the SOC prediction model to obtain the SOC prediction result, wherein the SOC prediction model is constructed based on a neural network and a battery electrochemical model.

[0041] The temperature data, voltage data, and current data are input into the SOH prediction model to obtain the SOH prediction result. The SOH prediction model is constructed based on support vector regression optimized by particle swarm optimization.

[0042] Based on the SOC prediction results and the SOH prediction results, the operating status of the industrial and commercial energy storage device is obtained.

[0043] The beneficial effects of this invention are as follows: By preprocessing and correcting the collected data, this invention can accurately and effectively collect data from energy storage devices. On the one hand, by combining an electrochemical model with a neural network, this invention improves the generalization ability of the model and the accuracy of SOC prediction; on the other hand, by using support vector regression optimized by particle swarm optimization to predict SOH, it can improve the accuracy of SOH prediction. Therefore, by combining the state of charge (SOC) and state of health (SOH), this invention can accurately monitor the operating status of industrial and commercial energy storage devices. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a framework diagram of a working status monitoring system for an industrial and commercial energy storage device according to an embodiment of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0048] Example 1:

[0049] In power supply systems, operators often cannot directly observe the current condition of batteries from the outside, such as whether the current and voltage are stable, whether the temperature is normal, or whether the power is too high. These simple parameters can be monitored through sensing. To effectively improve the power supply quality of batteries, it is often necessary to master indicators that more intuitively and deeply reflect the dynamic performance and wear status of batteries. Among these, the State of Charge (SOC) and State of Health (SOH) are the most important. SOC is a crucial operating parameter reflecting the battery's current capacity, while SOH is an indicator of battery wear and aging. Battery aging manifests in many ways. Therefore, this invention monitors SOC and SOH to monitor the operating status of commercial and industrial energy storage devices.

[0050] like Figure 1 As shown, the present invention discloses a working status monitoring system for industrial and commercial energy storage devices, including: a data acquisition module, a data transmission module, a SOC prediction module, a SOH prediction module, and a status monitoring module;

[0051] The data acquisition module is used to acquire monitoring data, which includes real-time temperature data, initial voltage data, and initial current data of industrial and commercial energy storage devices, and to perform correction processing on the initial voltage data and initial current data.

[0052] The data transmission module is used to package and compress temperature data, processed voltage data, and current data, and then transmit the packaged and compressed data to the status monitoring module.

[0053] The SOC prediction module is used to input voltage and current data into the SOC prediction model to obtain the SOC prediction result. The SOC prediction model is built based on neural networks and battery electrochemical models.

[0054] The SOH prediction module is used to input temperature data, voltage data, and current data into the SOH prediction model to obtain the SOH prediction results. The SOH prediction model is constructed based on support vector regression optimized by particle swarm optimization.

[0055] The status monitoring module is used to obtain the operating status of industrial and commercial energy storage devices based on the SOC and SOH prediction results.

[0056] Furthermore, the data acquisition module includes a calibration model building unit and a data calibration unit;

[0057] The calibration model construction unit is used to acquire historical voltage and current data, and input the initial voltage calibration model and initial current calibration model to calculate the voltage calibration parameters and current calibration parameters, and acquire the voltage calibration model and current calibration model; the data calibration unit is used to calibrate the real-time acquired initial voltage and initial current data using the voltage calibration model and current calibration model respectively.

[0058] Furthermore, obtaining historical voltage and current data includes:

[0059] Initial historical voltage and current data are collected within a preset time period using a preset sampling rate. Missing values ​​are filled and outliers are replaced to obtain historical voltage and current data. For missing and outliers, the mean of the two adjacent data points is used for filling and replacement, respectively.

[0060] The method for correcting the initial voltage data acquired in real time using a voltage correction model is as follows:

[0061] U = [(U' / c) – b] × a

[0062] Where U is the corrected voltage value, U' is the voltage value acquired before correction, a is the voltage amplification factor, b is the first DC bias, c is the reference source factor, c = C' / C, where C is the reference source voltage and C' is the reference voltage read by the system;

[0063] The method for correcting the real-time acquired initial current data using a current correction model is as follows:

[0064] I = [(I' / c–d)] × e

[0065] Where I is the corrected current value, I' is the current value collected before correction, e is the current amplification factor, and d is the second DC bias.

[0066] In this embodiment, in the data correction unit, multiple sets of real U, U', and C' data are collected and substituted into the voltage correction model to calculate voltage calibration parameters a and b, which are then stored in the system FLASH. Similarly, multiple sets of real I, I', and C' data are collected and substituted into the current correction model to calculate current calibration parameters d and e, which are also stored in the system FLASH. This embodiment, by correcting the collected data, solves the problem of large deviations in data acquisition accuracy, greatly improving the accuracy of energy storage data acquisition.

[0067] Furthermore, the data transmission module includes a data packaging and compression unit and a transmission unit; the data packaging and compression unit is used to package and compress the processed voltage and current data in the form of email; the transmission unit is used to provide email services to the platform-side communication PC, CDMA device, CDMA gateway, and Internet, wherein the platform-side communication PC, CDMA device, CDMA gateway, and Internet are wirelessly connected in sequence.

[0068] Furthermore, the SOC prediction module includes: a model fitting unit, a data processing unit, and an SOC prediction unit;

[0069] The model fitting unit is used to fit the battery electrochemical model using a neural network to obtain the discrete equation of the SOC prediction model.

[0070] The data processing unit is used to process current data, voltage data, and discrete equations using an improved unscented Kalman filter algorithm to obtain the Kalman gain.

[0071] The SOC prediction unit is used to estimate the parameters of commercial and industrial energy storage devices at the current moment based on the Kalman gain, and obtain the SOC prediction result at the current moment.

[0072] Furthermore, fitting the battery electrochemical model using neural networks includes: fitting the nonlinear discrete-time state-space model of the battery electrochemical model using a BP neural network.

[0073]

[0074] Where SOC(k+1) represents the SOC at the next time step, Vt(k+1) represents the RC dynamic voltage at the next time step, SOC(k) represents the SOC at the current time step, Vt(k) represents the RC dynamic voltage at the current time step, and T... S The sampling time is represented by C0, the discharge coefficient by Vout(k), the output voltage at the current moment by Voc(SOC(k)), the open-circuit voltage by Voc(SOC(k)), and the input current by I(k). In the neural network fitting formula: θ represents the fitting result of θ1, θ2, or θ3, s represents the number of neurons in the hidden layer, I represents the current current, U represents the current voltage, T represents the current temperature, and ω represents the current voltage. i1 ω i2 ω i3 These represent the input layer weights, ω and ω, respectively. i Indicates the output layer weights, b i b represents the input layer threshold, and b represents the output layer threshold.

[0075] A high-order neural network was used to fit the SOC-Voc curve of the battery electrochemical model:

[0076]

[0077] Where, N che and N try P represents the number of neurons in the hidden layer. n (SOC (k) ) represents the Chebyshev polynomial, ω sin ω represents the weights of the Chebyshev polynomial. sin ω represents the weights of the trigonometric polynomial. cos SOC represents the weights of a trigonometric polynomial. (k) This indicates the current SOC.

[0078] Specifically, the improved unscented Kalman filter algorithm is used to process current and voltage data, as well as discrete equations, including:

[0079] S1: Initialize the error covariance matrix;

[0080] S2: The error covariance matrix is ​​decomposed using the SVD decomposition method to obtain the Sigma sampling points;

[0081] S3: Initialize system noise Q;

[0082] S4: Time update, calculate the mean and variance of the state variables based on the Sigma point and the system noise Q;

[0083] S5: Measurement update, calculate the mean and variance of the output variable based on the Sigma point and system noise Q; calculate the Kalman gain based on the mean and variance of the state variable and the mean and variance of the output variable;

[0084] S6: Filter update, update the error covariance matrix based on the variance of the state variable, the variance of the output variable, the measured value of the output variable and the Kalman gain, and return to step S1;

[0085] S7: Update the system noise Q using the Sage-Husa estimator and return to step S3.

[0086] The filter update includes: when voltage data is missing, the error covariance matrix is ​​stopped from being updated, and the error covariance matrix from the previous time step is directly reused.

[0087] In SVR regression model construction, model parameters significantly impact regression accuracy. The penalty factor C adjusts the algorithm's tolerance for error during learning. If set too small, errors caused by out-of-interval samples decrease, leading to poorer accuracy in the trained regression model. Conversely, a large C improves accuracy but reduces generalization ability. ε determines the interval between the two boundaries of the regression interval; excessively large or small ε values ​​also affect accuracy and generalization, and influence outlier handling, thus influencing the algorithm's robustness. To improve accuracy and generalization, this embodiment employs Particle Swarm Optimization (PSO) to optimize the parameters in the estimation model. The kernel function parameter determines the distribution of sample data in the high-dimensional feature space, influencing the complexity of the learning process. Particle Optimization (PSO) is a biomimetic-based global optimization method. Its basic idea is to treat each particle as an optimal individual to solve the optimization problem, and to simulate it according to biomimetic principles. In seeking the optimal solution, each particle searches based on its own and other particles' "flight experience." A fitness function is set to measure the current position, and the particle updates its velocity and positioning using its own extreme points and the global extreme points.

[0088] Furthermore, the SOH prediction module includes an SOH prediction model building unit;

[0089] The SOH prediction model building unit is used to train the SVR model based on the PSO algorithm, update the individual extreme values ​​and global extreme values ​​of the particle swarm, determine the optimal values ​​of the parameters, and build an SVR regression model as the SOH prediction model based on the optimal values.

[0090] Furthermore, training the SVR model based on the PSO algorithm and updating the individual and global extrema of the particle swarm includes:

[0091] S1. Use the mean squared error of the SVR regression model as the fitness function, and initialize the particle swarm position and velocity;

[0092] S2. Train the SVR regression model using sample data, calculate particle fitness, and obtain the globally optimal initial value based on the results;

[0093] S3. Update the particle's position and velocity;

[0094] S4. Train the SVR regression model and update the individual extrema and global extrema of the particles;

[0095] S5. Determine if the termination condition has been met. If the number of iterations reaches the set threshold, the current global extreme value is used as the optimization output; otherwise, return to S3 to continue the iteration.

[0096] Example 2:

[0097] A method for monitoring the operational status of an industrial and commercial energy storage device includes:

[0098] Real-time acquisition of temperature data, initial voltage data, and initial current data of industrial and commercial energy storage devices, and correction processing of initial voltage data and initial current data;

[0099] Temperature data, processed voltage data, and current data are packaged and compressed, and the packaged and compressed data is transmitted to the SOC prediction model and the SOH prediction model.

[0100] Voltage and current data are input into the SOC prediction model to obtain the SOC prediction results. The SOC prediction model is built based on neural networks and battery electrochemical models.

[0101] Temperature, voltage, and current data are input into the SOH prediction model to obtain the SOH prediction results. The SOH prediction model is constructed based on support vector regression optimized by particle swarm optimization.

[0102] Based on the SOC and SOH prediction results, the operating status of industrial and commercial energy storage devices is obtained.

[0103] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A monitoring system for the operational status of an industrial and commercial energy storage device, characterized in that, include: Data acquisition module, data transmission module, SOC prediction module, SOH prediction module, status monitoring module; The data acquisition module is used to acquire monitoring data, wherein the monitoring data is real-time acquired temperature data, initial voltage data, and initial current data of industrial and commercial energy storage devices, and to perform correction processing on the initial voltage data and the initial current data. The data acquisition module includes a calibration model construction unit and a data calibration unit; The correction model construction unit is used to acquire historical voltage data and historical current data, and input the initial voltage correction model and the initial current correction model to calculate the voltage correction parameters and current correction parameters, and to obtain the voltage correction model and the current correction model. The data correction unit is used to correct the initial voltage data and initial current data acquired in real time using the voltage correction model and the current correction model, respectively. The method for correcting the real-time acquired initial voltage data using the voltage correction model is as follows: U = [(U' / c) – b] × a; Where U is the corrected voltage value, U' is the voltage value acquired before correction, a is the voltage amplification factor, b is the first DC bias, c is the reference source factor, c = C' / C, where C is the reference source voltage and C' is the reference voltage read by the system; The method for correcting the real-time acquired initial current data using the current correction model is as follows: I = [(I' / c–d)] × e; Where I is the corrected current value, I' is the current value collected before correction, e is the current amplification factor, and d is the second DC bias; The data transmission module is used to package and compress temperature data, processed voltage data and current data, and transmit the packaged and compressed data to the status monitoring module. The SOC prediction module is used to input the voltage data and the current data into the SOC prediction model to obtain the SOC prediction result, wherein the SOC prediction model is constructed based on a neural network and a battery electrochemical model. The SOH prediction module is used to input the temperature data, voltage data and current data into the SOH prediction model to obtain the SOH prediction result. The SOH prediction model is constructed based on support vector regression optimized by particle swarm optimization. The status monitoring module is used to obtain the operating status of the industrial and commercial energy storage device based on the SOC prediction result and the SOH prediction result.

2. The industrial and commercial energy storage device operating status monitoring system according to claim 1, characterized in that, Obtaining historical voltage and current data includes: Initial historical voltage data and initial historical current data are collected within a preset time period by a preset sampling rate. Missing values ​​are filled and outliers are replaced in the initial historical voltage data and initial historical current data to obtain the historical voltage data and the historical current data. For missing values ​​and outliers, the mean of the two adjacent data points of the missing value and the outlier are used for filling and replacement respectively.

3. The industrial and commercial energy storage device operating status monitoring system according to claim 1, characterized in that, The SOC prediction module includes: a model fitting unit, a data processing unit, and an SOC prediction unit; The model fitting unit is used to fit the battery electrochemical model using a neural network to obtain the discrete equation of the SOC prediction model. The data processing unit is used to process the current data, the voltage data, and the discrete equations using an improved unscented Kalman filter algorithm to obtain the Kalman gain. The SOC prediction unit is used to estimate the parameters of the industrial and commercial energy storage device at the current moment based on the Kalman gain, and obtain the SOC prediction result at the current moment.

4. The industrial and commercial energy storage device operating status monitoring system according to claim 3, characterized in that, Fitting battery electrochemical models using neural networks includes: fitting the nonlinear discrete-time state-space model of the battery electrochemical model using a BP neural network; Using high-order neural networks for battery electrochemical models The curve is fitted.

5. The industrial and commercial energy storage device operating status monitoring system according to claim 1, characterized in that, The SOH prediction module includes an SOH prediction model construction unit; The SOH prediction model construction unit is used to train the SVR model based on the PSO algorithm and update the individual extreme values ​​and global extreme values ​​of the particle swarm, determine the optimal values ​​of the parameters, and construct an SVR regression model based on the optimal values ​​as the SOH prediction model.

6. The industrial and commercial energy storage device operating status monitoring system according to claim 5, characterized in that, Training the SVR model based on the PSO algorithm and updating the individual and global maxima of the particle swarm includes: S1. Use the mean squared error of the SVR regression model as the fitness function, and initialize the particle swarm position and velocity; S2. Train the SVR regression model using sample data, calculate particle fitness, and obtain the globally optimal initial value based on the results; S3. Update the position and velocity of the particles; S4. Train the SVR regression model and update the individual extreme values ​​and global extreme values ​​of the particles; S5. Determine if the termination condition has been met. If the number of iterations reaches the set threshold, the current global extreme value is used as the optimization output; otherwise, return to S3 to continue the iteration.

7. A method for monitoring the operating status of an industrial and commercial energy storage device, characterized in that, include: Real-time acquisition of temperature data, initial voltage data, and initial current data from industrial and commercial energy storage devices, and correction processing of the initial voltage data and the initial current data, including: Acquire historical voltage and current data, and input them into the initial voltage correction model and initial current correction model to calculate the voltage correction parameters and current correction parameters, and obtain the voltage correction model and current correction model; The voltage correction model and the current correction model are used to correct the initial voltage data and initial current data acquired in real time, respectively. The method for correcting the real-time acquired initial voltage data using the voltage correction model is as follows: U = [(U' / c) – b] × a; Where U is the corrected voltage value, U' is the voltage value acquired before correction, a is the voltage amplification factor, b is the first DC bias, c is the reference source factor, c = C' / C, where C is the reference source voltage and C' is the reference voltage read by the system; The method for correcting the real-time acquired initial current data using the current correction model is as follows: I = [(I' / c–d)] × e; Where I is the corrected current value, I' is the current value collected before correction, e is the current amplification factor, and d is the second DC bias; The temperature data, processed voltage data, and current data are packaged and compressed, and the packaged and compressed data is transmitted to the SOC prediction model and the SOH prediction model. The voltage data and the current data are input into the SOC prediction model to obtain the SOC prediction result, wherein the SOC prediction model is constructed based on a neural network and a battery electrochemical model. The temperature data, voltage data, and current data are input into the SOH prediction model to obtain the SOH prediction result. The SOH prediction model is constructed based on support vector regression optimized by particle swarm optimization. Based on the SOC prediction results and the SOH prediction results, the operating status of the industrial and commercial energy storage device is obtained.

Images