Method and system for monitoring running state of sodium ion battery in full life cycle

Abstract

The invention provides a method and system for monitoring the running state of a sodium ion battery in the full life cycle, and the method comprises the steps: setting a plurality of branch layers in a model, enabling each branch layer to undertake different prediction tasks, receiving the output from a hidden layer as the input, and carrying out the data processing through the respective full-connection neurons and activation functions; in the neural network, input data is gradually transmitted to an output layer from an input layer; an output result of the output layer is obtained through the activation function and weight calculation between the layers; completing forward propagation, and comparing a prediction result with a corresponding real label value to obtain a loss function value so as to quantify the difference between the prediction result and an actual label; the error is reversely propagated to the whole network by using a loss function, parameter loss contribution is solved, and the gradient is reversely calculated; minimizing the loss function according to the gradient direction; gradient information is solved; and adjusting neural network parameters. The technical problem that the accuracy and the real-time performance of battery state monitoring are poor is solved.

Description

Technical Field

[0001] The present invention relates to the field of new energy monitoring, and in particular to a method and system for monitoring the operating status of a sodium ion battery throughout its life cycle. Background Art

[0002] Sodium-ion batteries, as a key technology for electrochemical energy storage, achieve energy storage through the reversible migration of sodium ions between positive and negative electrodes. They are similar to lithium-ion batteries but at a lower cost. They have a simple manufacturing process, abundant raw materials, long life, and can operate stably over a wide temperature range. In addition, sodium-ion batteries have high-rate charge and discharge capabilities, can quickly respond to power needs, and are suitable for electric vehicles, smart grids and other fields. In the energy storage market, sodium-ion batteries are seen as a substitute for lead-acid batteries, complementing lithium-ion batteries to meet the diverse needs of high performance, low cost and safety. They have shown great potential in the fields of renewable energy integration and grid energy storage, providing support for the transformation of the energy structure.

[0003] However, over time, and under the continuous influence of different usage conditions and environmental factors, including temperature, humidity, and charge and discharge rates, the performance of sodium-ion batteries will inevitably decline gradually. This performance degradation is not only directly reflected in the significant reduction in battery life, causing the device to perform poorly in situations where continuous power supply is required, but also potentially increases the risk of battery failure, which may cause a series of safety issues, including but not limited to overheating, short circuits, and even explosions. Therefore, in order to ensure the reliable operation of the equipment, avoid unexpected downtime or accidents caused by battery performance degradation, and optimize the battery usage strategy from an economic perspective and maximize its effective service life, it is particularly important to accurately and timely monitor the electrical performance degradation state of sodium-ion batteries.

[0004] For example, the existing invention patent application document "A method and system for predicting the working health status of batteries in energy storage power stations" with publication number CN114547849A, the existing method includes: obtaining the historical operation data of the battery in the energy storage power station and preprocessing the original data; according to the characteristics of the battery, intercepting the platform voltage of the battery, using this as a reference to obtain the battery charging platform interval voltage and current data for a period of time from the charging segment of the historical operation data, and obtaining the charging capacity value of each time through ampere-hour integration to obtain the original battery health status change sequence; using the empirical mode decomposition algorithm to separate the battery cycle change segment and the linear decay segment from the original battery health status change sequence; respectively establishing data-driven models for the cycle change segment and the linear decay segment to achieve the change trend prediction of the battery linear decay segment and the seasonal cycle segment. However, the aforementioned prior art is mainly based on the battery charging data for modeling, so the accuracy of the existing solution for battery monitoring needs to be further improved.

[0005] The existing invention patent application document "Sodium-ion battery life detection method, system, electronic device and storage medium" with publication number CN118483597A, the existing method includes: obtaining reconstructed characterization data; obtaining the charge prediction of the sodium-ion battery based on multi-step iterative prediction; constructing the optimal separation hyperplane in the high-dimensional feature space; using the continuous scaling prediction method to separate and predict the energy feature set; by selecting a preset threshold to control the characteristic feature variables after screening and sorting, to obtain the correlation between the single-cell cycle life of the sodium-ion battery and the reconstructed characterization data; based on the reconstructed characterization data, obtaining the battery parameter model of the sodium-ion battery, generating a prediction relationship between the battery parameter model and the SOH prediction situation, and generating the cascade life prediction of the sodium-ion battery based on the correlation and the prediction relationship. However, the multi-step iterative prediction method adopted by the existing scheme is more technically difficult than the single-step iterative prediction method, and the applicable scenario of the existing technology is the long-term prediction of sodium-ion batteries, which can capture the complex dynamic changes of batteries over a period of time. Only in the application scenario of single-step prediction, because it does not involve the cumulative error of the predicted value, can the accuracy be guaranteed. In the long-term iterative calculation, as time goes by, the errors of the initial prediction will accumulate in the later stage, resulting in a decrease in the accuracy of long-term prediction.

[0006] The existing invention patent application document "Battery Remaining Life Prediction Method and Device Based on Sodium Ion Battery" with publication number CN118566743A, the existing method includes: obtaining target observation data of the target sodium ion battery; determining multiple time steps according to the target observation data, and generating a target long time series according to the multiple time steps; inputting the target long time series into the battery life prediction model to obtain the predicted battery remaining life of the target sodium ion battery, wherein the battery life prediction model includes a dynamic self-attention encoding layer and a temporal attention decoding layer, the dynamic self-attention encoding layer includes a first multi-head self-attention mechanism and a first feedforward network, the first feedforward network includes a first fully connected layer, a second fully connected layer and an activation function layer, and the temporal attention decoding layer includes a second multi-head self-attention mechanism and a second feedforward network. However, the self-attention mechanism adopted in the existing scheme allows the model to directly refer to other positions of the entire sequence when calculating the representation of a certain position in the sequence, so as to capture long-distance dependencies. However, the structure of the self-attention mechanism itself is relatively complex, requiring more parameters and computing resources, which is easy to increase the training difficulty and overfitting risk of the model.

[0007] The existing invention patent application document "Sodium-ion battery cloud SOH estimation and remaining life prediction method, device and program" with publication number CN116754960A, the existing method includes: step 1, collecting operating data of sodium-ion batteries in use in the cloud, and processing the operating data to obtain different battery operating segments; step 2, for each battery operating segment corresponding to different voltage intervals including full charging and discharging and non-full charging and discharging, extracting health factors related to the battery health status from the operating data and establishing a health factor library; wherein the health factors specifically include three types of parameters for characterizing the battery health status, namely thermodynamics, kinetics and statistics; step 3, screening the health factors in the health factor library using at least one screening and evaluation system, Get a subset of characteristic parameters that are strongly correlated with SOH; Step 4: Take the characteristic parameter subset as input, and the battery SOH as the estimated target output, and use the battery operation data that has been used for a certain period of time and is in a stable decay state to train the machine learning algorithm, and reflect the mapping relationship between the characteristic parameters and the battery SOH through the trained algorithm; Step 5: Take the characteristic parameter subset and the corresponding battery SOH trajectory as input, and the battery RUL as the predicted target output, and train the deep learning algorithm, and reflect the mapping relationship between the characteristic parameters, battery SOH and remaining life through the trained algorithm; Step 6: Use the trained machine learning algorithm and deep learning algorithm, input the characteristic parameters collected in real time to obtain the SOH estimation value and RUL prediction value. In step 3 of the existing scheme, the health factor screening uses one or a combination of two or three of the filtering, wrapping and embedded screening and evaluation systems. Although this method is simple and direct, it requires a large search space and high computational cost, and cannot quickly and real-time deliver characteristic parameters.

[0008] In summary, the prior art has technical problems of poor accuracy and real-time performance of battery status monitoring. Summary of the invention

[0009] The technical problem to be solved by the present invention is: how to solve the technical problem of poor accuracy and real-time performance of battery status monitoring in the prior art.

[0010] The present invention adopts the following technical solutions to solve the above technical problems: The method for monitoring the operating status of a sodium ion battery during its entire life cycle comprises:

[0011] S1. Under the preset energy storage scenario, obtain the sodium-ion battery charge and discharge operating curve, use the preset equivalent circuit model to simulate and diagnose the battery cells in the cycle charge and discharge test, identify the battery performance change information in a specific cycle, and output the identification results;

[0012] S2. Performing a cyclic charge and discharge test on the sodium ion battery to collect key state parameters of the sodium ion battery;

[0013] S3, preprocessing the key state parameters, constructing the battery operation state feature vector, and constructing the performance attenuation database of the sodium ion battery in the energy storage condition according to the recognition result and the battery operation state feature vector;

[0014] S4. According to the energy storage operating condition performance attenuation database, a convolutional neural network model is trained to obtain a suitable convolutional neural network model, and the operating status of the battery cell is monitored using the suitable convolutional neural network model based on the input feature vector.

[0015] The present invention monitors the electrical performance attenuation state of sodium ion batteries, which can not only help to promptly discover early signs of battery performance degradation and take corresponding measures for maintenance or replacement, but also provide valuable data support for the design and optimization of battery management systems, thereby improving the overall operating efficiency, safety and service life of the equipment.

[0016] The present invention utilizes a convolutional neural network deep learning algorithm. Compared with the feature parameter extraction algorithm adopted in the above-mentioned existing invention patents, it is a machine learning algorithm that can provide the best accuracy in many use cases. It has powerful feature extraction capabilities, the advantages of parameter sharing and sparse connections, translation invariance and visual interpretability, etc., and uses a trained applicable convolutional neural network model to monitor the operating status of the battery cell in real time. Through multi-layer convolution operations, the network can quickly and automatically learn the local structure and global characteristics of the real-time state of the battery, and cooperate with the equivalent circuit model to accurately and real-time reflect the real state of the battery through offline training and online parameter updating, thereby realizing online evaluation of the operating status of the sodium ion battery, in order to at least solve the deficiencies in the above-mentioned prior art.

[0017] In a more specific technical solution, the identification results include: circuit model parameters and key indicators of operating status.

[0018] In a more specific technical solution, the convolutional neural network model in S4 includes: an input layer, a hidden layer, no less than 2 branch layers and an output layer. The input layer receives an input feature vector and passes the input feature vector to the hidden layer.

[0019] In a more specific technical solution, in the convolutional neural network model, each branch layer undertakes the difference prediction task, uses the branch layer to receive the output features of the hidden layer, and processes the output features of the hidden layer through the fully connected neurons and activation functions of each branch layer;

[0020] Using the following logic, the convolutional neural network model is expressed as the relationship between input A and output B according to the function F(A|Θ) composed of nonlinear mapping and preset learning parameters:

[0021] B=F(A|Θ)=f L{…f2[f1(X|θ1)|θ2]|θ L}

[0022] In the formula, f l (·|θ l ) is the calculation process of hidden layer and branch layer, Θ is the set of parameters θ.

[0023] The present invention cooperates with the peak-shaving and frequency-regulating dispatching requirements of the power grid for the energy storage system, takes into account the operating status of the battery cell during the entire life cycle under the corresponding energy storage conditions, and adopts an equivalent circuit model and a neural network model to evaluate the battery status. It can accurately and real-time reflect the true status of the battery, and realize online evaluation of the operating status of the sodium-ion battery through offline training and online parameter updating, providing an accurate basis for battery maintenance.

[0024] In a more specific technical solution, in S4, a convolutional neural network model is used to convert the voltage, current, and temperature values ​​[V1, I1, T1, V2, I2, T2, …, V n0 ,I n0 ,T n0 ] is mapped to the current battery health state SOH.

[0025] The trained neural network model in the present invention is imported, which defines in detail the network hierarchy, the number of neurons in each layer, the type of activation function, and the connection weights and biases between layers, providing a basis for executing computing tasks.

[0026] In a more specific technical solution, in S4, a convolutional neural network model is used for forward propagation, and the input data is gradually transferred from the input layer to the output layer, and the output result of the output layer is obtained by the activation function and the weight calculation between each layer;

[0027] The prediction results of the convolutional neural network model are compared with the corresponding true label values, and the loss function value is calculated; according to the value of the loss function, the difference between the prediction results and the actual labels is quantified.

[0028] In a more specific technical solution, the loss function is used to back-propagate the error to the entire convolutional neural network model, by calculating the contribution of each parameter to the loss and using the chain rule to reversely calculate the gradient from the output layer to the input layer; according to the direction of the aforementioned gradient, the value of each parameter is updated to minimize the loss function; based on the calculated gradient information, the parameters in the neural network are adjusted accordingly to obtain a suitable convolutional neural network model.

[0029] In a more specific technical solution, in S4, a suitable convolutional neural network model is imported to define the network hierarchy, the number of neurons in each layer, the activation function type, and the inter-layer connection weights and biases.

[0030] In a more specific technical solution, in S4, the input feature vector is sent to the applicable convolutional neural network model, and each neuron in the applicable convolutional neural network model obtains a weighted sum result by weighted summing the output from the previous layer, and applies an activation function to introduce nonlinear characteristics and realize feature transformation;

[0031] The activation function is used to perform nonlinear mapping on the weighted summation results to obtain and output the prediction results.

[0032] In the convolutional neural network model adopted by the present invention, each neuron performs weighted summation of the output from the previous layer and applies an activation function to introduce nonlinear characteristics and realize feature transformation. The role of the activation function is to perform nonlinear mapping on the result of the weighted summation, thereby enhancing the expression ability and adaptability of the model and finally outputting the prediction result.

[0033] In a more specific technical solution, the operating status monitoring system of sodium-ion batteries throughout their life cycle includes:

[0034] The energy storage scenario data acquisition module is used to obtain the sodium-ion battery charge and discharge condition curve under the preset energy storage scenario, use the preset equivalent circuit model to simulate and diagnose the battery cells in the cycle charge and discharge test, identify the battery performance change information in a specific cycle, and output the identification results;

[0035] The state parameter acquisition module performs a cyclic charge and discharge test on the sodium ion battery to collect key state parameters of the sodium ion battery;

[0036] The performance attenuation database construction module is used to pre-process the key state parameters, construct the battery operation state feature vector, and construct the performance attenuation database of the sodium-ion battery in the energy storage condition according to the recognition results and the battery operation state feature vector. The performance attenuation database construction module is connected with the state parameter acquisition module and the energy storage scenario data acquisition module;

[0037] The battery cell status monitoring module is used to train the convolutional neural network model according to the energy storage operating condition performance attenuation database to obtain a suitable convolutional neural network model, obtain and monitor the battery cell operating status using the suitable convolutional neural network model based on the input feature vector, and the battery cell status monitoring module is connected to the performance attenuation database construction module.

[0038] Compared with the prior art, the present invention has the following advantages:

[0039] The present invention monitors the electrical performance attenuation state of sodium ion batteries, which can not only help to promptly discover early signs of battery performance degradation and take corresponding measures for maintenance or replacement, but also provide valuable data support for the design and optimization of battery management systems, thereby improving the overall operating efficiency, safety and service life of the equipment.

[0040] The present invention cooperates with the peak-shaving and frequency-regulating dispatching requirements of the power grid for the energy storage system, takes into account the operating status of the battery cell during the entire life cycle under the corresponding energy storage conditions, and adopts an equivalent circuit model and a neural network model to evaluate the battery status. It can accurately and real-time reflect the true status of the battery, and realize online evaluation of the operating status of the sodium-ion battery through offline training and online parameter updating, providing an accurate basis for battery maintenance.

[0041] The trained neural network model in the present invention is imported, which defines in detail the network hierarchy, the number of neurons in each layer, the type of activation function, and the connection weights and biases between layers, providing a basis for executing computing tasks.

[0042] In the convolutional neural network model adopted by the present invention, each neuron performs weighted summation of the output from the previous layer and applies an activation function to introduce nonlinear characteristics and realize feature transformation. The role of the activation function is to perform nonlinear mapping on the result of the weighted summation, thereby enhancing the expression ability and adaptability of the model and finally outputting the prediction result.

[0043] The present invention solves the technical problem of poor accuracy and real-time performance of battery status monitoring in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 This is a schematic diagram of the basic steps of the method for monitoring the operating status of a sodium ion battery during its entire life cycle according to Example 1 of the present invention;

[0045] Figure 2 Schematic diagram of SOH prediction using a convolutional neural network model according to Example 2 of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in combination with the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Example 1

[0048] like Figure 1As shown in the figure, the operating status monitoring method of sodium ion batteries during their entire life cycle includes the following basic steps:

[0049] S1. Use the sodium-ion battery charge and discharge operating curve suitable for energy storage scenarios, use the equivalent circuit model to simulate and diagnose the battery cells in the cycle charge and discharge test, identify long-term battery performance changes, and output results including circuit model parameters and key operating status indicators;

[0050] S2. In the above cycle charge and discharge test, key parameters of the sodium ion battery state are collected;

[0051] S3, pre-processing the collected voltage, current and other parameters, constructing the battery operation state feature vector, and forming a performance attenuation database of sodium ion batteries under energy storage conditions;

[0052] S4. According to the performance attenuation database formed by the battery operation status feature vector, a convolutional neural network algorithm model is trained, and the trained convolutional neural network algorithm model is used to monitor the battery cell operation status according to the input feature vector.

[0053] In this embodiment, the key parameters of the sodium ion battery state include but are not limited to: voltage, current, internal resistance and SOH;

[0054] like Figure 2 As shown, in this embodiment, a convolutional neural network model is constructed, in which no less than 2 branch layers are set, each branch layer undertakes a different prediction task, and the aforementioned branch layer receives the output from the hidden layer as input, and performs data processing through respective fully connected neurons and activation functions;

[0055] In this embodiment, the architecture of the aforementioned convolutional neural network model includes but is not limited to: an input layer, a hidden layer, no less than 2 branch layers and an output layer. Specifically, the input layer receives feature data and passes the feature data to the hidden layer.

[0056] In this embodiment, assuming that the input and output of the convolutional neural network model are A and B respectively, the relationship between the input and output can be expressed as the following formula through a function F(A|Θ) consisting of a nonlinear mapping and a learning parameter:

[0057] B=F(A|Θ)=f L {…f2[f1(X|θ1)|θ2]|θ L}

[0058] Among them, f l (·|θ l ) is the calculation process of hidden layer and branch layer, Θ is the set of parameters θ.

[0059] In this embodiment, after analyzing the measured data of the sodium ion battery, it is found that the voltage, current, and temperature curves during the cycle are relatively stable because the cycle mode is a fixed constant current-constant voltage mode, and as the sodium ion battery ages, there is a corresponding relationship with the capacity of the sodium ion battery, that is, the voltage, current, and temperature during a constant current-constant voltage charging process are all one-dimensional time series data that change with time and correspond to the same SOH. The characteristic of the sodium ion battery that the data in this local range corresponds to the same feature and the weights are consistent. Therefore, the convolutional neural network algorithm model can convert the voltage, current, and temperature values ​​[V1, I1, T1, V2, I2, T2, …, V n0 ,I n0 ,T n0 ] is mapped to the current battery health state SOH.

[0060] The convolutional neural network model is used for forward propagation, and the input data is gradually transferred from the input layer to the output layer. The output result of the output layer is obtained through the activation function and the weight calculation between each layer.

[0061] The prediction results of the convolutional neural network model are compared with the corresponding true label values, and the value of the loss function is calculated; according to the value of the loss function, the difference between the prediction results and the actual labels is quantified.

[0062] The error is back-propagated to the entire convolutional neural network model using the loss function. The contribution of each parameter to the loss is calculated, and the gradient is calculated backwards from the output layer to the input layer using the chain rule. According to the direction of the aforementioned gradient, the value of each parameter is updated to minimize the loss function. Based on the calculated gradient information, the parameters in the neural network are adjusted accordingly.

[0063] In this embodiment, a trained convolutional neural network model is imported. In the convolutional neural network model, the network hierarchy, the number of neurons in each layer, the activation function type, and the connection weights and biases between layers are defined in detail, providing a basis for executing computing tasks.

[0064] Input data is fed into the network, starting from the input layer, and the data flows through each neuron in turn. After being processed layer by layer, information is continuously transmitted until it reaches the final output layer. In the above process, each neuron performs a weighted summation of the output from the previous layer and applies an activation function to introduce nonlinear characteristics and achieve feature transformation. The role of the activation function is to perform a nonlinear mapping on the result of the weighted summation, thereby enhancing the expressiveness and adaptability of the model and outputting the prediction result.

[0065] In summary, the present invention monitors the electrical performance attenuation state of sodium-ion batteries, which can not only help to promptly discover the early signs of battery performance degradation and take corresponding measures for maintenance or replacement, but also provide valuable data support for the design and optimization of battery management systems, thereby improving the overall operating efficiency, safety and battery life of the equipment.

[0066] The present invention cooperates with the peak-shaving and frequency-regulating dispatching requirements of the power grid for the energy storage system, takes into account the operating status of the battery cell during the entire life cycle under the corresponding energy storage conditions, and adopts an equivalent circuit model and a neural network model to evaluate the battery status. It can accurately and real-time reflect the true status of the battery, and realize online evaluation of the operating status of the sodium-ion battery through offline training and online parameter updating, providing an accurate basis for battery maintenance.

[0067] The trained neural network model in the present invention is imported, which defines in detail the network hierarchy, the number of neurons in each layer, the type of activation function, and the connection weights and biases between layers, providing a basis for executing computing tasks.

[0068] In the convolutional neural network model adopted by the present invention, each neuron performs weighted summation of the output from the previous layer and applies an activation function to introduce nonlinear characteristics and realize feature transformation. The role of the activation function is to perform nonlinear mapping on the result of the weighted summation, thereby enhancing the expression ability and adaptability of the model and finally outputting the prediction result.

[0069] The present invention solves the technical problem of poor accuracy and real-time performance of battery status monitoring in the prior art.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for monitoring the operating status of a sodium ion battery during its entire life cycle, characterized in that: The method comprises: S1. Under the preset energy storage scenario, obtain the sodium-ion battery charge and discharge operating curve, use the preset equivalent circuit model to simulate and diagnose the battery cells in the cycle charge and discharge test, identify the battery performance change information in a specific cycle, and output the identification results; S2. Performing a cyclic charge and discharge test on the sodium ion battery to collect key state parameters of the sodium ion battery; S3, preprocessing the key state parameters, constructing a battery operation state feature vector, and constructing a performance attenuation database of the sodium ion battery in energy storage conditions according to the recognition result and the battery operation state feature vector; S4. According to the energy storage operating condition performance attenuation database, a convolutional neural network model is trained to obtain a suitable convolutional neural network model, and the operating status of the battery cell is monitored using the suitable convolutional neural network model based on the input feature vector.

2. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 1, characterized in that: The identification results include: circuit model parameters and key indicators of operating status.

3. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 1, characterized in that: The convolutional neural network model in S4 includes: an input layer, a hidden layer, no less than 2 branch layers and an output layer. The input layer is used to receive the input feature vector and pass the input feature vector to the hidden layer.

4. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 3, characterized in that: In the convolutional neural network model, each of the branch layers undertakes the difference prediction task, uses the branch layer to receive the output features of the hidden layer, and processes the output features of the hidden layer through the fully connected neurons and activation functions of each of the branch layers; Using the following logic, the convolutional neural network model is expressed as the relationship between input A and output B according to the function F(A|Θ) composed of nonlinear mapping and preset learning parameters: B=F(A|Θ)=f L {…f2[f1(X|θ1)|θ2]|θ L } In the formula, f l (·|θ l ) is the calculation process of hidden layer and branch layer, Θ is the set of parameters θ.

5. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 1, characterized in that: In S4, the convolutional neural network model is used to calculate the voltage, current, and temperature values ​​[V1, I1, T1, V2, I2, T2, ..., V n0 ,I n0 ,T n0 ] is mapped to the current battery health state SOH.

6. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 1, characterized in that: In S4, a convolutional neural network model is used for forward propagation, and the input data is gradually transferred from the input layer to the output layer, and the output result of the output layer is obtained by calculating the activation function and the weights between the layers; The prediction results of the convolutional neural network model are compared with the corresponding true label values, and the loss function value is calculated; according to the value of the loss function, the difference between the prediction results and the actual labels is quantified.

7. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 6, characterized in that: The error is back-propagated to the entire convolutional neural network model using the loss function, and the contribution of each parameter to the loss is calculated, and the gradient is reversely calculated from the output layer to the input layer using the chain rule; the value of each parameter is updated according to the direction of the aforementioned gradient to minimize the loss function; and the parameters in the neural network are adjusted accordingly based on the calculated gradient information to obtain the applicable convolutional neural network model.

8. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 1, characterized in that: In S4, the applicable convolutional neural network model is imported to define the network hierarchy, the number of neurons in each layer, the activation function type, and the inter-layer connection weights and biases.

9. The method for monitoring the operating status of a sodium ion battery during its entire life cycle according to claim 1, characterized in that: In S4, the input feature vector is sent to an applicable convolutional neural network model, and each neuron in the applicable convolutional neural network model obtains a weighted sum result by performing weighted summation on the output from the previous layer, and applies an activation function to introduce nonlinear characteristics and realize feature transformation; The activation function is used to perform nonlinear mapping on the weighted summation result to obtain and output a prediction result.

10. The operating status monitoring system of sodium ion battery in the whole life cycle is characterized by: The system comprises: The energy storage scenario data acquisition module is used to obtain the sodium-ion battery charge and discharge condition curve under the preset energy storage scenario, use the preset equivalent circuit model to simulate and diagnose the battery cells in the cycle charge and discharge test, identify the battery performance change information in a specific cycle, and output the identification results; A state parameter acquisition module performs a cyclic charge and discharge test on the sodium ion battery to acquire key state parameters of the sodium ion battery; A performance attenuation database construction module is used to pre-process the key state parameters, construct a battery operation state feature vector, and construct a performance attenuation database of the sodium ion battery in energy storage conditions according to the recognition result and the battery operation state feature vector. The performance attenuation database construction module is connected to the state parameter acquisition module and the energy storage scenario data acquisition module; A cell state monitoring module is used to train a convolutional neural network model according to the energy storage operating condition performance attenuation database to obtain a suitable convolutional neural network model, obtain and monitor the cell operating state using the suitable convolutional neural network model based on the input feature vector, and the cell state monitoring module is connected to the performance attenuation database construction module.

Images