Battery state estimation method and device, electronic equipment and storage medium

Abstract

The invention provides a battery state estimation method and device, electronic equipment and a storage medium, and the method comprises the steps: implanting an optical fiber temperature sensing device into a lithium battery, and collecting the battery data and temperature data of the lithium battery in which the optical fiber temperature sensing device is implanted in the charging and discharging process; a second-order RC equivalent circuit model of the lithium battery is constructed, parameter identification is carried out on the second-order RC equivalent circuit model based on the least square method and the battery data, and a parameter identification result is determined; the parameter identification result and the battery data are processed based on a Kalman filter, and the open-circuit voltage of the lithium battery is determined; and performing battery health state estimation processing on the open-circuit voltage, the real-time current and the temperature data based on a battery state estimation model, and outputting the battery health state of the lithium battery implanted into the sensing device. The method has the advantages of high sensing precision, strong real-time performance, abundant state estimation items and strong anti-interference performance, and can realize real-time high-precision battery state estimation and management.

Description

Technical Field

[0001] The present application relates to the field of battery technology, and in particular to a battery state estimation method, device, electronic device, and storage medium. Background Art

[0002] Lithium batteries may face safety hazards and performance degradation during operation. Therefore, accurate prediction of the state of charge (SOC) and state of health (SOH) of lithium batteries is of great significance for their performance optimization and safe operation. Current research focuses on optimizing algorithm models for battery state estimation. However, traditional monitoring methods often rely on limited external sensors, which make it difficult to fully reflect the thermal distribution and electrochemical state inside the battery, resulting in low accuracy and poor real-time performance of battery parameter perception. In addition, BMS usually uses complex algorithm models to estimate the battery state. These models have high requirements for data volume and data accuracy, and lack an accurate description of the complex dynamic processes inside the battery. Especially in the battery aging stage, the accuracy of the model decreases significantly, resulting in inaccurate state estimation. Therefore, how to accurately estimate the battery state has become a technical issue that cannot be underestimated. Summary of the Invention

[0003] In view of this, the purpose of this application is to provide a battery state estimation method, device, electronic device and storage medium, which have the advantages of high perception accuracy, strong real-time performance, rich state estimation items and strong anti-interference ability, and can realize real-time and high-precision battery state estimation and management.

[0004] An embodiment of the present application provides a battery state estimation method, the battery state estimation method comprising:

[0005] Implanting an optical fiber temperature sensing device inside a lithium battery to collect battery data and temperature data of the lithium battery implanted with the sensing device during the charging and discharging process;

[0006] Constructing a second-order RC equivalent circuit model of the lithium battery implanted in the sensing device, performing parameter identification on the second-order RC equivalent circuit model based on a least squares method and the battery data, and determining a parameter identification result;

[0007] Processing the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device;

[0008] The battery health status is estimated based on the open circuit voltage, the real-time current in the battery data, and the temperature data, and the battery health status of the lithium battery implanted in the sensing device is output; wherein, the battery state estimation model is obtained by iteratively training a convolutional neural network model.

[0009] In one possible implementation, performing parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data to determine the parameter identification result includes:

[0010] Discretizing the second-order RC equivalent circuit model based on a Laplace transform algorithm to determine a discretized form of the second-order RC equivalent circuit model;

[0011] Processing the discretized form based on the least squares method and the battery data to determine an initial parameter identification result of the second-order RC equivalent circuit model;

[0012] The initial parameter identification result is updated based on the least squares method at each time step to determine the parameter identification result.

[0013] In one possible implementation, the processing of the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device includes:

[0014] Performing state prediction processing, covariance prediction processing, Kalman gain calculation processing, and correction processing on the parameter identification result and the battery data based on the Kalman filter, and outputting a target charge state of the lithium battery implanted in the sensing device;

[0015] Based on the relationship curve between the charge state and the open circuit, the open circuit voltage corresponding to the target charge state is determined.

[0016] In one possible implementation, performing battery health state estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on a battery state estimation model, and outputting the battery health state of the lithium battery implanted in the sensing device, includes:

[0017] performing abnormal data removal and data normalization processing on the open circuit voltage, the real-time current, and the temperature data to determine health characterization data;

[0018] The battery health state estimation process is performed on the health characterization data based on the convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer of the battery state estimation model, and the battery health state is output.

[0019] In one possible implementation, the convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer based on the battery state estimation model perform battery health state estimation processing on the health characterization data and output the battery health state, including:

[0020] performing feature extraction on the health representation data to determine a feature vector of the health representation data;

[0021] Performing convolution processing on the feature vector based on multiple filters in a convolution layer, and outputting the feature vector after convolution of each filter;

[0022] Performing pooling processing on each convolved feature vector based on the pooling layer, and outputting multiple pooled feature vectors;

[0023] Performing one-dimensional data expansion processing on the plurality of pooled feature vectors based on the smoothing layer, and outputting a plurality of one-dimensional feature vectors;

[0024] Based on the long short-term memory network layer and the random drop layer, the long-term and short-term dependency capture processing is performed on the multiple one-dimensional feature vectors, the hidden features at each time step are determined, the hidden features at each time step are fully connected and normalized, and the battery health status is output.

[0025] In a possible implementation manner, the battery state estimation model is determined by the following steps:

[0026] Input the sample open circuit voltage, sample real-time current and sample temperature data of multiple sample lithium batteries into the convolutional neural network model for battery health state estimation processing, and output the predicted battery health state of each sample lithium battery;

[0027] Processing the actual battery health state and the predicted battery health state of each sample lithium battery based on a mean square error loss function to determine a loss value of the convolutional neural network model;

[0028] If the loss value is greater than or equal to the preset loss value, the network parameters of the convolutional neural network model are changed and training continues; if the loss value is less than the preset loss value, the convolutional neural network model is used as the battery state estimation model.

[0029] In a possible implementation, the optical fiber temperature sensing device includes a multi-point fiber Bragg grating sensor.

[0030] The present application also provides a battery state estimation device, which includes:

[0031] A data acquisition module is used to implant an optical fiber temperature sensing device inside a lithium battery and collect battery data and temperature data of the lithium battery implanted with the sensing device during the charging and discharging process;

[0032] an equivalent circuit model construction module, configured to construct a second-order RC equivalent circuit model of the lithium battery implanted in the sensing device, perform parameter identification on the second-order RC equivalent circuit model based on a least squares method and the battery data, and determine a parameter identification result;

[0033] an open circuit voltage determination module, configured to process the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device;

[0034] A battery health status determination module is used to perform battery health status estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on a battery status estimation model, and output the battery health status of the lithium battery implanted in the sensing device; wherein the battery status estimation model is obtained by iteratively training a convolutional neural network model.

[0035] An embodiment of the present application also provides an electronic device, comprising: a processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and when the machine-readable instructions are executed by the processor, the steps of the battery status estimation method as described above are performed.

[0036] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the battery state estimation method as described above are executed.

[0037] The present invention provides a battery state estimation method, device, electronic device, and storage medium. The method comprises: implanting an optical fiber temperature sensing device inside a lithium battery to collect battery data and temperature data of the lithium battery implanted with the sensing device during charging and discharging; constructing a second-order RC equivalent circuit model of the lithium battery implanted with the sensing device, performing parameter identification based on the least squares method and the second-order RC equivalent circuit model of the battery data to determine a parameter identification result; processing the parameter identification result and the battery data using a Kalman filter to determine the open circuit voltage of the lithium battery implanted with the sensing device; and performing battery health state estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on the battery state estimation model to output the battery health state of the lithium battery implanted with the sensing device. The battery state estimation model is obtained by iteratively training a convolutional neural network model. The method has the advantages of high perception accuracy, strong real-time performance, rich state estimation items, and strong anti-interference ability, and can achieve real-time and high-precision battery state estimation and management.

[0038] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0040] Figure 1 A flowchart of a battery status estimation method provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of a battery state estimation method provided in an embodiment of the present application;

[0042] Figure 3 This is one of the structural diagrams of a battery state estimation device provided in an embodiment of the present application;

[0043] Figure 4 This is a second structural diagram of a battery state estimation device provided in an embodiment of the present application;

[0044] Figure 5 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The components of the embodiments of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, each other embodiment obtained by those skilled in the art without making creative work falls within the scope of protection of the present application.

[0046] First, the application scenarios to which this application is applicable are introduced. This application can be applied in the field of battery technology.

[0047] Research has found that lithium batteries may face safety hazards and performance degradation during operation. Therefore, accurate prediction of the state of charge (SOC) and state of health (SOH) of lithium batteries is of great significance for their performance optimization and safe operation. Current research focuses on optimizing algorithm models for battery state estimation. However, traditional monitoring methods often rely on limited external sensors, which make it difficult to fully reflect the thermal distribution and electrochemical state inside the battery, resulting in low accuracy and poor real-time performance of battery parameter perception. In addition, BMS usually uses complex algorithm models to estimate the battery state. These models have high requirements for data volume and data accuracy, and lack an accurate description of the complex dynamic processes inside the battery. Especially in the battery aging stage, the accuracy of the model decreases significantly, resulting in inaccurate state estimation. Therefore, how to accurately estimate the battery state has become a technical issue that cannot be underestimated.

[0048] Based on this, an embodiment of the present application provides a battery status estimation method, which has the advantages of high perception accuracy, strong real-time performance, rich status estimation items, and strong anti-interference ability, and can realize real-time and high-precision battery status estimation and management.

[0049] Please read Figure 1 , Figure 1 This is a flow chart of a battery state estimation method provided in an embodiment of the present application. Figure 1 As shown in , the battery state estimation method provided by the embodiment of the present application includes:

[0050] S101: Implanting an optical fiber temperature sensing device inside a lithium battery to collect battery data and temperature data of the lithium battery implanted with the sensing device during the charging and discharging process.

[0051] In this step, the cylindrical lithium battery to be monitored is used as a physical entity, and a fiber-optic temperature sensing device is implanted. Quasi-distributed temperature self-sensing is achieved through embedded design within the smart battery. Charge and discharge cycles are performed on the lithium battery with the fiber-optic temperature sensing device. The battery management system collects and uploads data such as battery temperature and current to the workstation.

[0052] Preferably, the optical fiber temperature sensing device includes a multi-point Fiber Bragg Grating (FBG) sensor.

[0053] Here, the implementation process of implanting a fiber optic temperature sensing device inside a lithium battery is as follows: First, in an inert gas environment, use a tube cutter to open the positive terminal of the battery, and use Kapton tape to isolate the positive and negative poles to prevent short circuits. Next, a fiber optic sensor with a polytetrafluoroethylene tube (PTFE) is implanted into the isolation layer of the battery, and the positive pole of the battery is resealed with an epoxy resin engineering adhesive to protect its interior. After resealing, the battery is left to stand for 24 hours, and after the engineering adhesive dries, Kapton tape is used to provide additional protection for the engineering adhesive. After the lithium battery preparation is completed, the lithium battery is placed on the test bench, the battery is connected to the battery cycler through a copper busbar, and the FBG sensor is connected to the optical fiber demodulator to achieve real-time self-sensing of the internal temperature of the lithium battery.

[0054] S102: Constructing a second-order RC equivalent circuit model of the lithium battery implanted in the sensing device, performing parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data, and determining a parameter identification result.

[0055] In this step, the battery management system is a second-order RC equivalent circuit model of the lithium battery implanted with the sensing device. The battery management system uses the least squares method to perform parameter identification on the second-order RC equivalent circuit model based on the collected battery data to determine the parameter identification results.

[0056] The parameter identification results include the polarization internal resistance of the RC loop, the polarization capacitance of the RC loop, and the ohmic internal resistance.

[0057] Here, the battery management system establishes a second-order RC equivalent circuit model for the lithium battery, and its mathematical expression is:

[0058]

[0059] V t (t) = V oc (t)+V p1 (t)+V p2 (t)+R0(t)I(t)

[0060] Where I is the load current of the battery, R1 and R2 are the polarization internal resistance of the RC loop, C1 and C2 are the polarization capacitance of the RC loop, R0 is the ohmic internal resistance, V p1 and V p2 is the voltage across the two RC loops, V oc is the open circuit voltage, V t is the terminal voltage, and t is the time point.

[0061] In one possible implementation, performing parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data to determine the parameter identification result includes:

[0062] A: Discretize the second-order RC equivalent circuit model based on a Laplace transform algorithm to determine a discretized form of the second-order RC equivalent circuit model.

[0063] Here, the discrete form after Laplace transform is:

[0064]

[0065] θ e =[k1 k2 k3 k4 k5] T

[0066]

[0067] Among them, K is the time point, θ e is a parameter matrix, which includes parameter information such as the polarization internal resistance of the RC loop, the polarization capacitance of the RC loop, and the ohmic internal resistance. is a regression vector consisting of the calculable historical current and polarization voltage states.

[0068] in:

[0069]

[0070] Among them, R1 and R2 are the polarized internal resistances of the RC loop, C1 and C2 are the polarized capacitances of the RC loop, R0 is the ohmic internal resistance, k1, k2, k3, k4, and k5 correspond to R0, R1, C1, R2, and C2, respectively, and T is the sampling period.

[0071] B: Processing the discretized form based on the least squares method and the battery data to determine an initial parameter identification result of the second-order RC equivalent circuit model.

[0072] Here, the parameter matrix θ e It can be identified by the least squares method shown below:

[0073]

[0074] Where P(k) is the covariance matrix at time point k; K(k) is the gain matrix at time point k; e(k) is the estimated error at time point k; y is the output variable; and I is the identity matrix.

[0075] C: updating the initial parameter identification result based on the least squares method at each time step to determine the parameter identification result.

[0076] S103: Processing the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device.

[0077] In this step, the parameter identification results and battery data are processed using a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device.

[0078] In one possible implementation, the processing of the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device includes:

[0079] a: Based on the Kalman filter, the parameter identification result and the battery data are subjected to state prediction processing, covariance prediction processing, Kalman gain calculation processing and correction processing, and the target charge state of the lithium battery implanted in the sensing device is output.

[0080] Here, the parameter identification results and battery data are subjected to state prediction processing, covariance prediction processing, Kalman gain calculation processing and correction processing according to the Kalman filter, and the target charge state of the lithium battery implanted in the sensing device is output.

[0081] First, initialize Then, two steps are performed: prediction and update. The prediction step includes state prediction and covariance prediction, and the update step mainly includes Kalman gain and posterior state correction.

[0082] Here, the mathematical expression of state prediction is as follows:

[0083]

[0084] in, is the state estimation at the current time k, A(k) is the state transfer matrix at the current time k, and the state estimation includes the charge state at the current time k and the polarization voltage on the two RC networks. is the initial state matrix, is the initial covariance estimation matrix.

[0085] Among them, the mathematical expression of covariance prediction is as follows:

[0086]

[0087] in, is the covariance estimate at the current time k, δ W is the input noise.

[0088] The EKF update step mainly includes Kalman gain and posterior state correction, where the mathematical expression of Kalman gain is as follows:

[0089]

[0090] Among them, L(k) is the Kalman gain, C(k) is the observation matrix, δ v is the observation noise.

[0091] Here, the mathematical expression of the posterior state correction is as follows:

[0092]

[0093] in, for The linear equation of .

[0094] Here, the covariance correction is finally performed on the posterior error, and its mathematical expression is as follows:

[0095]

[0096] b: Based on the relationship curve between the charge state and the open circuit, determine the open circuit voltage corresponding to the target charge state.

[0097] Here, the open circuit voltage corresponding to the target charge state is determined according to the relationship curve between the charge state and the open circuit.

[0098] S104: Perform battery health status estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on a battery status estimation model, and output the battery health status of the lithium battery implanted in the sensing device; wherein, the battery status estimation model is obtained by iteratively training a convolutional neural network model.

[0099] In this step, the battery health status is estimated by using the battery state estimation model on the open circuit voltage, hourly current and temperature data, and the battery health status is output.

[0100] Battery health (SOH) is an important indicator for assessing battery aging or degradation, reflecting the battery's health. When defined in terms of capacity decay, SOH is the ratio of the battery's current remaining capacity to its rated capacity.

[0101]

[0102] Among them, C n is the current remaining capacity of the battery, C0 is the rated capacity of the battery, and SOH is calculated based on the acquired data.

[0103] In one possible implementation, performing battery health state estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on a battery state estimation model, and outputting the battery health state of the lithium battery implanted in the sensing device, includes:

[0104] i: performing abnormal data removal and data normalization processing on the open circuit voltage, the real-time current, and the temperature data to determine health characterization data.

[0105] Here, the maximum and minimum normalization algorithm is used to normalize the open circuit voltage, real-time current and temperature data.

[0106] ii: Based on the convolution layer, pooling layer, smoothing layer, long short-term memory network layer and random dropout layer of the battery state estimation model, the health characterization data is processed to estimate the battery health state and output the battery health state.

[0107] Here, the convolution layer, pooling layer, smoothing layer, long short-term memory network layer and random dropout layer are used to estimate the battery health status of the health representation data and output the battery health status.

[0108] In one possible implementation, the convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer based on the battery state estimation model perform battery health state estimation processing on the health characterization data and output the battery health state, including:

[0109] (1): Extract features from the health characterization data to determine a feature vector of the health characterization data.

[0110] (2): Convolution processing is performed on the feature vector based on multiple filters in the convolution layer, and the feature vector after convolution of each filter is output.

[0111] Here, the battery state estimation model is set up with two convolutional layers and one pooling layer to perform convolution training on the data. The convolutional layer in the convolution block contains 6 filters, each filter contains 6 convolution kernels, the convolution kernel size is 6×1, and the ReLU function is used as the activation function of the convolution layer.

[0112] Here, for a convolutional layer, the convolution calculation process is as follows:

[0113]

[0114] in, is the feature vector after the i-th convolution of the l-th layer; f is the activation function; W i l is the weight matrix of the i-th filter in the l-th layer; * is the convolution operator symbol; Xl-1 is the output of the l-1th layer; is the i-th bias from the l-1th layer to the lth layer.

[0115] (3): Based on the pooling layer, pooling processing is performed on each convolutional feature vector, and multiple pooled feature vectors are output.

[0116] Here, the pooling calculation process is as follows:

[0117]

[0118] in, is the element in the i-th feature vector of the l+1-th layer after pooling; is the element in the i-th feature vector of the l-th layer after pooling; D j is the area covered by the jth pooling.

[0119] (4): Based on the smoothing layer, one-dimensional data expansion processing is performed on the multiple pooled feature vectors to output multiple one-dimensional feature vectors.

[0120] (5): Based on the long short-term memory network layer and the random drop layer, the long-term and short-term dependency capture processing of the multiple one-dimensional feature vectors is performed to determine the hidden features at each time step, and the hidden features at each time step are fully connected and normalized to output the battery health status.

[0121] Here, the LSTM network processes input data through mechanisms such as forget gate, input gate, unit state candidate, unit state update, output gate and hidden state unit, thereby effectively capturing long-term and short-term dependencies in the time series. In this LSTM structure, the number of hidden layer neurons is set to 48. In order to simplify the complexity of the model and improve the ability to resist noise, a random dropout layer (Dropout Layer) is introduced, and its random dropout rate is set to 0.2 to avoid overfitting and improve the generalization ability of the model. For an LSTMN neural unit, it is defined that at each time step t, the hidden state h t From the data x at the same time step t Update, the hidden state h of the previous time step t-1 , input gate i t , input node g t , forget gate f t , output gate o t and storage unit c t , the update formula is:

[0122]

[0123] Where W and b are the layer weights and biases respectively: Wfx and W fh They are respectively the forget gate relative to x t and h t-1 The two weight matrices b f is the bias term, σ is the Sigmoid function; similarly, W ix and W ih They are respectively the input gate relative to x t and h t-1 The two weight matrices b i is the bias term; W gx and W gh They are respectively relative to x in the tanh activation function t and h t-1 The two weight matrices b g is the bias term; W ox and W oh They are the output gate relative to x t and h t-1 The two weight matrices b o is the bias term and t is the time step.

[0124] Preferably, through the dropout layer, some hidden outputs are randomly masked so that these neurons do not affect the forward propagation during training. The estimated result of the standard open circuit voltage matrix is ​​output based on the following formula:

[0125] pred i =relu(dropout((W out h t +b out ),1))

[0126] Among them, W out and b out They are the discard layer weight and bias, and the output value pred i is the SOH estimation result corresponding to the input feature i, relu is selected as the activation function, h t is the hidden state at time step t.

[0127] In a possible implementation manner, the battery state estimation model is determined by the following steps:

[0128] I: Input the sample open circuit voltage, sample real-time current, and sample temperature data of multiple sample lithium batteries into the convolutional neural network model for battery health status estimation processing, and output the predicted battery health status of each sample lithium battery.

[0129] Here, the process of the convolutional neural network model estimating the battery health status of sample information is consistent with the processing process of the battery status estimation model, and will not be repeated in this section.

[0130] II: Based on the mean square error loss function, the actual battery health state and the predicted battery health state of each sample lithium battery are processed to determine the loss value of the convolutional neural network model.

[0131] Here, the loss value is determined by the following formula:

[0132]

[0133] Among them, MSE is the loss value, n is the number of training samples, i is the true value, pred i is the model predicted value.

[0134] III: If the loss value is greater than or equal to the preset loss value, the network parameters of the convolutional neural network model are changed and training continues; if the loss value is less than the preset loss value, the convolutional neural network model is used as the battery state estimation model.

[0135] For further information, see Figure 2 , Figure 2 This is a schematic diagram of the battery state estimation method provided in the embodiment of the present application. Figure 2 As shown, a fiber optic temperature sensing device is implanted inside a lithium battery. Battery and temperature data from the implanted battery during the charging and discharging process are collected. A second-order RC equivalent circuit model of the implanted battery is constructed. A Kalman filter is used to process the parameter identification results of the second-order RC equivalent circuit model and the battery data to determine the state of charge. Based on the curve of the charge state and the open circuit voltage, the corresponding open circuit voltage is determined. The open circuit voltage, real-time current, and temperature data are then input into a battery state estimation model to determine the battery health status.

[0136] This application constructs an intelligent battery cell with internal temperature sensing capabilities by implanting FBG sensors. On this basis, a battery equivalent circuit model is established, and the model's parameter identification is achieved using a least squares estimation algorithm. Furthermore, a deep fusion of the extended Kalman filter (EKF) and the CNN-LSTM neural network is used to achieve multi-scale parameter joint estimation of the battery's state of charge (SOC) and state of health (SOH) taking thermal parameters into account. This method has the advantages of high perception accuracy, strong real-time performance, rich state estimation items, and strong anti-interference capabilities, enabling real-time and high-precision battery state estimation and management.

[0137] An embodiment of the present application provides a battery state estimation method, comprising: implanting an optical fiber temperature sensing device within a lithium battery to collect battery data and temperature data from the lithium battery implanted with the sensing device during charging and discharging; constructing a second-order RC equivalent circuit model of the lithium battery implanted with the sensing device, performing parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data to determine a parameter identification result; processing the parameter identification result and the battery data using a Kalman filter to determine the open circuit voltage of the lithium battery implanted with the sensing device; and performing battery health state estimation processing on the open circuit voltage, real-time current in the battery data, and temperature data based on the battery state estimation model to output the battery health state of the lithium battery implanted with the sensing device. The battery state estimation model is obtained by iteratively training a convolutional neural network model. The method has the advantages of high perception accuracy, strong real-time performance, rich state estimation items, and strong anti-interference capabilities, and can achieve real-time and high-precision battery state estimation and management.

[0138] See also Figure 3 、 Figure 4 , Figure 3 This is one of the structural diagrams of a battery state estimation device provided in an embodiment of the present application. Figure 4 This is a second structural diagram of a battery state estimation device provided in an embodiment of the present application. Figure 3 As shown in , the battery state estimation device 300 includes:

[0139] The data acquisition module 310 is used to implant an optical fiber temperature sensing device inside the lithium battery and collect battery data and temperature data of the lithium battery implanted with the sensing device during the charging and discharging process;

[0140] an equivalent circuit model construction module 320 for constructing a second-order RC equivalent circuit model of the lithium battery implanted in the sensing device, performing parameter identification on the second-order RC equivalent circuit model based on a least squares method and the battery data, and determining a parameter identification result;

[0141] an open circuit voltage determination module 330 for processing the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device;

[0142] The battery health status determination module 340 is used to perform battery health status estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on the battery status estimation model, and output the battery health status of the lithium battery implanted in the sensing device; wherein, the battery status estimation model is obtained by iteratively training a convolutional neural network model.

[0143] Furthermore, when the equivalent circuit model construction module 320 is used to perform parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data to determine the parameter identification result, the equivalent circuit model construction module 320 is specifically used to:

[0144] Discretizing the second-order RC equivalent circuit model based on a Laplace transform algorithm to determine a discretized form of the second-order RC equivalent circuit model;

[0145] Processing the discretized form based on the least squares method and the battery data to determine an initial parameter identification result of the second-order RC equivalent circuit model;

[0146] The initial parameter identification result is updated based on the least squares method at each time step to determine the parameter identification result.

[0147] Furthermore, when the open circuit voltage determination module 330 is used to process the parameter identification result and the battery data based on the Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device, the open circuit voltage determination module 330 is specifically used to:

[0148] Performing state prediction processing, covariance prediction processing, Kalman gain calculation processing, and correction processing on the parameter identification result and the battery data based on the Kalman filter, and outputting a target charge state of the lithium battery implanted in the sensing device;

[0149] Based on the relationship curve between the charge state and the open circuit, the open circuit voltage corresponding to the target charge state is determined.

[0150] Furthermore, when the battery health status determination module 340 is used to perform battery health status estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on the battery state estimation model and output the battery health status of the lithium battery implanted in the sensing device, the battery health status determination module 340 is specifically used to:

[0151] performing abnormal data removal and data normalization processing on the open circuit voltage, the real-time current, and the temperature data to determine health characterization data;

[0152] The battery health state estimation process is performed on the health characterization data based on the convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer of the battery state estimation model, and the battery health state is output.

[0153] Furthermore, when the battery health state determination module 340 performs battery health state estimation processing on the health characterization data using the convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer based on the battery state estimation model and outputs the battery health state, the battery health state determination module 340 is specifically configured to:

[0154] performing abnormal data removal and data normalization processing on the open circuit voltage, the real-time current, and the temperature data to determine health characterization data;

[0155] performing feature extraction on the health representation data to determine a feature vector of the health representation data;

[0156] Performing convolution processing on the feature vector based on multiple filters in a convolution layer, and outputting the feature vector after convolution of each filter;

[0157] Performing pooling processing on each convolved feature vector based on the pooling layer, and outputting multiple pooled feature vectors;

[0158] Performing one-dimensional data expansion processing on the plurality of pooled feature vectors based on the smoothing layer, and outputting a plurality of one-dimensional feature vectors;

[0159] Based on the long short-term memory network layer and the random drop layer, the long-term and short-term dependency capture processing is performed on the multiple one-dimensional feature vectors, the hidden features at each time step are determined, the hidden features at each time step are fully connected and normalized, and the battery health status is output.

[0160] Further, such as Figure 4 As shown, the battery state estimation device 300 further includes a model training module 350. The model training module 350 determines the battery state estimation model through the following steps:

[0161] Input the sample open circuit voltage, sample real-time current and sample temperature data of multiple sample lithium batteries into the convolutional neural network model for battery health state estimation processing, and output the predicted battery health state of each sample lithium battery;

[0162] Processing the actual battery health state and the predicted battery health state of each sample lithium battery based on a mean square error loss function to determine a loss value of the convolutional neural network model;

[0163] If the loss value is greater than or equal to the preset loss value, the network parameters of the convolutional neural network model are changed and training continues; if the loss value is less than the preset loss value, the convolutional neural network model is used as the battery state estimation model.

[0164] The present invention provides a battery state estimation device, comprising: a data acquisition module for implanting an optical fiber temperature sensing device within a lithium battery to collect battery data and temperature data of the lithium battery implanted with the sensing device during charging and discharging; an equivalent circuit model construction module for constructing a second-order RC equivalent circuit model of the lithium battery implanted with the sensing device, performing parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data, and determining a parameter identification result; an open circuit voltage determination module for processing the parameter identification result and the battery data using a Kalman filter to determine the open circuit voltage of the lithium battery implanted with the sensing device; and a battery health status determination module for performing battery health status estimation processing on the open circuit voltage, real-time current in the battery data, and temperature data based on the battery state estimation model, and outputting the battery health status of the lithium battery implanted with the sensing device. The battery state estimation model is obtained by iteratively training a convolutional neural network model. The device has the advantages of high perception accuracy, strong real-time performance, rich state estimation items, and strong anti-interference ability, and can achieve real-time and high-precision battery state estimation and management.

[0165] See also Figure 5 , Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 5 As shown in FIG, the electronic device 500 includes a processor 510, a memory 520 and a bus 530.

[0166] The memory 520 stores machine-readable instructions executable by the processor 510. When the electronic device 500 is running, the processor 510 communicates with the memory 520 via the bus 530. When the machine-readable instructions are executed by the processor 510, the above-mentioned Figure 1 as well as Figure 2 The specific implementation of the steps of the battery state estimation method in the method embodiment shown can be found in the method embodiment, and will not be repeated here.

[0167] The embodiment of the present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the computer program can execute the above-mentioned Figure 1 as well as Figure 2 The specific implementation of the steps of the battery state estimation method in the method embodiment shown can be found in the method embodiment, and will not be repeated here.

[0168] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0169] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. There may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed may be through some communication interface, indirect coupling or communication connection of devices or units, which may be electrical, mechanical or other forms.

[0170] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0171] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0172] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that is executable by a processor. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0173] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The scope of protection of the present application is not limited thereto. Although the present application has been described in detail with reference to the above-mentioned embodiments, those skilled in the art should understand that any person skilled in the art can modify or easily conceive of changes to the technical solutions described in the above-mentioned embodiments within the technical scope disclosed in the present application, or perform equivalent replacements for some of the technical features thereof. These modifications, changes, or replacements do not deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A battery state estimation method, characterized in that: The battery state estimation method comprises: Implanting an optical fiber temperature sensing device inside a lithium battery to collect battery data and temperature data of the lithium battery implanted with the sensing device during the charging and discharging process; Constructing a second-order RC equivalent circuit model of the lithium battery implanted in the sensing device, performing parameter identification on the second-order RC equivalent circuit model based on a least squares method and the battery data, and determining a parameter identification result; Processing the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device; The battery health status is estimated based on the open circuit voltage, the real-time current in the battery data, and the temperature data, and the battery health status of the lithium battery implanted in the sensing device is output; wherein, the battery state estimation model is obtained by iteratively training a convolutional neural network model.

2. The battery state estimation method according to claim 1, characterized in that: The performing parameter identification on the second-order RC equivalent circuit model based on the least squares method and the battery data to determine the parameter identification result includes: Discretizing the second-order RC equivalent circuit model based on a Laplace transform algorithm to determine a discretized form of the second-order RC equivalent circuit model; Processing the discretized form based on the least squares method and the battery data to determine an initial parameter identification result of the second-order RC equivalent circuit model; The initial parameter identification result is updated based on the least squares method at each time step to determine the parameter identification result.

3. The battery state estimation method according to claim 1, wherein: The processing of the parameter identification result and the battery data based on the Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device includes: Performing state prediction processing, covariance prediction processing, Kalman gain calculation processing, and correction processing on the parameter identification result and the battery data based on the Kalman filter, and outputting a target charge state of the lithium battery implanted in the sensing device; Based on the relationship curve between the charge state and the open circuit, the open circuit voltage corresponding to the target charge state is determined.

4. The battery state estimation method according to claim 1, wherein: The method of performing battery health state estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on the battery state estimation model, and outputting the battery health state of the lithium battery implanted in the sensing device, includes: performing abnormal data removal and data normalization processing on the open circuit voltage, the real-time current, and the temperature data to determine health characterization data; The battery health state estimation process is performed on the health characterization data based on the convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer of the battery state estimation model, and the battery health state is output.

5. The battery state estimation method according to claim 4, characterized in that: The convolution layer, pooling layer, smoothing layer, long short-term memory network layer, and random dropout layer based on the battery state estimation model perform battery health state estimation processing on the health characterization data and output the battery health state, including: performing feature extraction on the health representation data to determine a feature vector of the health representation data; Performing convolution processing on the feature vector based on multiple filters in a convolution layer, and outputting the feature vector after convolution of each filter; Performing pooling processing on each convolved feature vector based on the pooling layer, and outputting multiple pooled feature vectors; Performing one-dimensional data expansion processing on the plurality of pooled feature vectors based on the smoothing layer, and outputting a plurality of one-dimensional feature vectors; Based on the long short-term memory network layer and the random drop layer, the long-term and short-term dependency capture processing is performed on the multiple one-dimensional feature vectors, the hidden features at each time step are determined, the hidden features at each time step are fully connected and normalized, and the battery health status is output.

6. The battery state estimation method according to claim 1, characterized in that: The battery state estimation model is determined by the following steps: Input the sample open circuit voltage, sample real-time current and sample temperature data of multiple sample lithium batteries into the convolutional neural network model for battery health state estimation processing, and output the predicted battery health state of each sample lithium battery; Processing the actual battery health state and the predicted battery health state of each sample lithium battery based on a mean square error loss function to determine a loss value of the convolutional neural network model; If the loss value is greater than or equal to the preset loss value, the network parameters of the convolutional neural network model are changed and training continues; if the loss value is less than the preset loss value, the convolutional neural network model is used as the battery state estimation model.

7. The battery state estimation method according to claim 1, characterized in that: The optical fiber temperature sensing device includes a multi-point Bragg fiber grating sensor.

8. A battery state estimation device, characterized in that: The battery state estimation device comprises: A data acquisition module is used to implant an optical fiber temperature sensing device inside a lithium battery and collect battery data and temperature data of the lithium battery implanted with the sensing device during the charging and discharging process; an equivalent circuit model construction module, configured to construct a second-order RC equivalent circuit model of the lithium battery implanted in the sensing device, perform parameter identification on the second-order RC equivalent circuit model based on a least squares method and the battery data, and determine a parameter identification result; an open circuit voltage determination module, configured to process the parameter identification result and the battery data based on a Kalman filter to determine the open circuit voltage of the lithium battery implanted in the sensing device; A battery health status determination module is used to perform battery health status estimation processing on the open circuit voltage, the real-time current in the battery data, and the temperature data based on a battery status estimation model, and output the battery health status of the lithium battery implanted in the sensing device; wherein the battery status estimation model is obtained by iteratively training a convolutional neural network model.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor. When the electronic device is running, the processor and the memory communicate through the bus, and the machine-readable instructions are executed by the processor to execute the steps of the battery state estimation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the battery state estimation method according to any one of claims 1 to 7 are executed.

Images