Charge state determination method and device, vehicle-mounted embedded system and computer equipment

Abstract

The invention discloses a state of charge determination method and device, a vehicle-mounted embedded system and computer equipment. The method comprises the steps that state data of a target battery are acquired, and the state data comprise the voltage, the current and the internal resistance of the target battery; constructing a target battery mechanism model corresponding to the target battery; based on the state data and a target battery mechanism model, determining an initial state-of-charge value of the target battery; and inputting the initial state-of-charge value into a target neural network model to obtain a target state-of-charge value of the target battery, the target neural network model being used for correcting the initial state-of-charge value. According to the invention, the technical problem that the error of estimating the state of charge by using the battery mechanism model is large is solved.

Description

Technical Field

[0001] The present invention relates to the field of battery technology, and in particular to a method and device for determining a state of charge, a vehicle-mounted embedded system, and a computer device. Background Art

[0002] In the context of the current global energy crisis and increasing environmental awareness, electric vehicles are gradually becoming a mainstream trend in the automotive industry as an effective way to reduce dependence on fossil fuels and reduce greenhouse gas emissions. Lithium-ion batteries, as the core energy storage component of electric vehicles, are directly related to the range, safety and overall cost of electric vehicles. Among them, the state of charge (SOC) of the battery is defined as the ratio between the remaining capacity and the total capacity of the battery, which is one of the most important battery status indicators that cannot be directly measured. Inaccurate SOC estimation may accelerate battery aging, reduce life, and increase user anxiety about the mileage of electric vehicles. In view of this, accurate battery SOC prediction can help improve battery operation safety and service life, thereby increasing the market penetration of electric vehicles in commercial terms.

[0003] At present, SOC estimation methods are mainly divided into two categories: mechanism model-based methods and data-driven methods. Models based on battery mechanisms can more accurately reflect the behavior of batteries under different working conditions, but they also have some limitations, such as the uncertainty of model parameters, the simplification of the model may not fully reflect the complex characteristics of the battery, and the inaccuracy and noise of sensor measurements, all of which may introduce errors in SOC estimation. Data-driven methods based on machine learning require a huge amount of data to train more accurate results and require a large amount of experimental data, which may be difficult to obtain in practical applications, especially for data requirements under various working conditions. In addition, the computing resources of vehicle-mounted embedded systems are limited, so it is impractical to train complex machine learning models directly on the vehicle.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a method and apparatus for determining a state of charge, a vehicle-mounted embedded system, and a computer device, so as to at least solve the technical problem that the error of estimating the state of charge using a battery mechanism model is large.

[0006] According to one aspect of an embodiment of the present invention, a method for determining a state of charge is provided, comprising: obtaining state data of a target battery, wherein the state data includes voltage, current, and internal resistance of the target battery; constructing a target battery mechanism model corresponding to the target battery; determining an initial state of charge value of the target battery based on the state data and the target battery mechanism model; inputting the initial state of charge value into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0007] Optionally, constructing a target battery mechanism model corresponding to the target battery includes: obtaining state data corresponding to the target battery under multiple working conditions; based on the target battery, selecting an initial battery mechanism model, wherein the initial battery mechanism model includes one of the following: an equivalent circuit model, an electrochemical model, and an empirical formula model; based on the state data corresponding to the target battery under multiple working conditions, optimizing the initial battery mechanism model to obtain the target battery mechanism model.

[0008] Optionally, when the target battery mechanism model is an equivalent circuit model, the initial state of charge value of the target battery is determined based on the state data and the target battery mechanism model, including: determining the state equation of the equivalent circuit model based on Kirchhoff's voltage law, wherein the state equation characterizes the state of the equivalent circuit model, and the state equation includes the state of charge; based on the Kalman filter algorithm, predicting and updating the state equation multiple times until a preset stop condition is reached to obtain the target state equation; and determining the initial state of charge value of the target battery based on the target state equation.

[0009] Optionally, before inputting the initial state of charge value into the target neural network model, the method includes: obtaining multiple initial state of charge values ​​determined based on the target battery mechanism model and multiple real state of charge values ​​obtained based on battery testing equipment; inputting the multiple initial state of charge values ​​and the multiple real state of charge values ​​as sample sets into the initial neural network model for training to obtain the target neural network model.

[0010] Optionally, the initial neural network model includes one of the following: Elman neural network, long short-term memory neural network, recursive neural network.

[0011] According to another aspect of an embodiment of the present invention, a vehicle-mounted embedded system is also provided, which applies any one of the above-mentioned state of charge determination methods, including: a mechanism model module, used to determine the initial state of charge value of the target battery; and an offline network module, used to correct the initial state of charge value based on a target neural network model to obtain a target state of charge value of the target battery.

[0012] According to another aspect of an embodiment of the present invention, a state of charge determination device is also provided, including: an acquisition module, used to acquire state data of a target battery, wherein the state data includes voltage, current and internal resistance of the target battery; a construction module, used to construct a target battery mechanism model corresponding to the target battery; a determination module, used to determine an initial state of charge value of the target battery based on the state data and the target battery mechanism model; and a correction module, used to input the initial state of charge value into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0013] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided, the non-volatile storage medium including a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute any one of the above-mentioned charge state determination methods.

[0014] According to another aspect of an embodiment of the present invention, a computer device is provided. The computer device includes a processor, and the processor is used to run a program. When the program is run, any one of the above-mentioned methods for determining the state of charge is executed.

[0015] According to yet another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, any one of the above-mentioned methods for determining a state of charge is implemented.

[0016] In an embodiment of the present invention, a state of charge determination method is adopted, by obtaining state data of a target battery, wherein the state data includes voltage, current and internal resistance of the target battery; constructing a target battery mechanism model corresponding to the target battery; determining an initial state of charge value of the target battery based on the state data and the target battery mechanism model; inputting the initial state of charge value into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value, thereby achieving the purpose of accurately estimating the state of charge of the battery, thereby realizing the technical effect of improving the accuracy and efficiency of the state of charge estimation, and further solving the technical problem of large errors in estimating the state of charge using a battery mechanism model. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a method for determining a state of charge is shown;

[0019] Figure 2 is a flow chart of a method for determining a state of charge according to an embodiment of the present invention;

[0020] Figure 3 is a curve diagram showing the relationship between vehicle speed, power and time under NEDC conditions provided according to an optional embodiment of the present invention;

[0021] Figure 4 is a circuit diagram of a second-order equivalent circuit model provided according to an optional embodiment of the present invention;

[0022] Figure 5 A flowchart of initial SOC value result correction of a mechanism model based on a long short-term memory neural network is provided according to an optional embodiment of the present invention;

[0023] Figure 6 is a curve diagram of the change of the loss function during the training process provided by an optional embodiment of the present invention;

[0024] Figure 7 is a flow chart of a method for estimating the state of charge of a lithium-ion battery provided according to an optional embodiment of the present invention;

[0025] Figure 8 4 is a structural block diagram of a device for determining a state of charge according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] According to an embodiment of the present invention, an embodiment of a method for determining a state of charge is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal for implementing a method for determining a state of charge. Figure 1 As shown, the computer terminal 10 may include one or more (102a, 102b, ..., 102n are used to illustrate) processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations are shown.

[0030] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0031] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the state of charge determination method in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the state of charge determination method of the above-mentioned application is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0032] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 .

[0033] Figure 2 is a flow chart of a method for determining a state of charge according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0034] Step S201, obtaining status data of a target battery, wherein the status data includes voltage, current and internal resistance of the target battery.

[0035] In this step, charge and discharge tests and pulse charge and discharge tests under dynamic conditions of lithium batteries can be carried out based on professional battery test equipment, such as a battery tester, to obtain the status data of the lithium battery, including battery system voltage, battery cell voltage, battery current, battery temperature, insulation resistance, cumulative charge and discharge capacity, cumulative charge and discharge energy, etc., wherein the charge and discharge test refers to fully charging the battery and then discharging it at a constant current or power until the battery voltage drops to a preset cut-off voltage, or until the battery is fully discharged. Considering that the battery is mostly in a multi-time-varying dynamic condition in the actual vehicle, it is possible to consider carrying out steady-state test conditions and dynamic test conditions, wherein the steady-state condition is a cross-current discharge condition, and the dynamic condition adopts the NEDC (New European Driving Cycle) condition, which simulates typical urban and suburban driving conditions to reflect the average behavior of the car in daily use. Figure 3 is a curve diagram showing the relationship between vehicle speed, power and time under NEDC conditions provided by an optional embodiment of the present invention, such as Figure 3 As shown, during the NEDC operating condition test, the corresponding relationship between vehicle speed and time is converted into the corresponding relationship between battery charging and discharging power and time through the whole vehicle model.

[0036] Step S202, constructing a target battery mechanism model corresponding to the target battery.

[0037] In this step, the mechanism model can be divided into several types, including equivalent circuit models, electrochemical models and empirical formula models. The equivalent circuit model includes a first-order equivalent circuit model, a second-order equivalent circuit model and a higher-order equivalent circuit model. The electrochemical model includes an SPM model, a P2D model and a heterogeneous model. Each model has its applicable scenarios and characteristics. In this embodiment, a second-order equivalent circuit model can be selected as the mechanism model because the equivalent circuit model can better balance the accuracy and computational complexity of the model. After selecting the mechanism model, the model is adjusted according to the test data to make the model as close to the physical entity as possible, thereby simulating the electrical characteristics, thermal characteristics and other state characteristics of the lithium battery, and realizing battery-related estimates as accurately as possible.

[0038] Step S203, determining an initial state of charge value of the target battery based on the state data and the target battery mechanism model.

[0039] In this step, the calculation of the initial state of charge is based on the state data and the mechanism model. Specifically, the mechanism model is first used to predict the state data such as the voltage, current and internal resistance of the battery in a stable state. Then, the predicted data is compared with the actual collected state data, and the model parameters are updated through the Kalman filter algorithm to minimize the prediction error. This process needs to be iterated until the model parameters and prediction errors reach a stable state. Through the iterative update of the mechanism model and the Kalman filter algorithm, a more accurate SOC estimate can be obtained. However, due to the defects of the mechanism model itself, there is still a large error with the true value. The determination of the initial state of charge value provides an important starting point for subsequent corrections, ensuring the continuity and accuracy of the entire estimation process.

[0040] Step S204, inputting the initial state of charge value into the target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0041] In this step, in order to correct the initial state of charge value, a target neural network model needs to be constructed first. The target neural network model is trained based on the estimated value of the battery mechanism model and the state of charge data obtained from the actual test. Its core purpose is to learn and map the difference function between the predicted value of the battery mechanism model and the actual SOC value. The steps of training the neural network algorithm through the training set data include setting the initial model, setting the initial parameters, model training, updating weights and biases, and obtaining the optimization model. After the training of the target neural network model is completed, the initial SOC value can be passed as input to the neural network. The target neural network model will correct the SOC according to the mapping function learned during the training process. Specifically, the target neural network model will adjust the SOC value to be closer to the actual state of charge of the battery based on its internal weight and bias parameters, combined with the historical state data of the battery. This process is essentially a data-driven refinement of the SOC estimation based on the battery mechanism model.

[0042] The present invention takes into account the inherent errors, sampling accuracy and noise influence of the mechanism-based battery model, and uses a neural network method to solve the errors that are difficult to find regularities. It also solves the problem that machine learning related algorithms are difficult to deploy in vehicle-mounted embedded systems. It chooses to train the neural network in an offline form to obtain a correction function, and directly couples the correction function to the mechanism model based on the equivalent circuit, which greatly reduces the demand for the amount of model training data, thereby greatly saving computing power, so that the feasibility of battery state of charge estimation is guaranteed while improving the accuracy of the estimation method.

[0043] Through the above steps, the purpose of accurately estimating the battery state of charge is achieved, thereby achieving the technical effect of improving the accuracy and efficiency of state of charge estimation, and further solving the technical problem of large errors in estimating the state of charge using a battery mechanism model.

[0044] As an optional embodiment, constructing a target battery mechanism model corresponding to the target battery includes: obtaining state data corresponding to the target battery under multiple working conditions; based on the target battery, selecting an initial battery mechanism model, wherein the initial battery mechanism model includes one of the following: an equivalent circuit model, an electrochemical model, and an empirical formula model; based on the state data corresponding to the target battery under multiple working conditions, optimizing the initial battery mechanism model to obtain the target battery mechanism model.

[0045] Optionally, the target lithium-ion battery needs to be tested on professional battery testing equipment to collect its status data under different working conditions. These working conditions include but are not limited to static, dynamic, different temperatures, different charge and discharge rates, etc., to cover various possible working conditions of the battery in actual applications. The status data should include key parameters such as the battery's voltage (V), current (I), temperature (T) and internal resistance (R).

[0046] After collecting the state data of the target battery, you can select an initial battery mechanism model, such as an equivalent circuit model, an electrochemical model, and an empirical formula model, as the basis for subsequent model optimization. The equivalent circuit model simulates the dynamic response of the battery through circuit components (such as RC links), while the electrochemical model considers the internal chemical reaction process of the battery more deeply, and the empirical formula model is a mathematical expression obtained by fitting the measured data. The specific model to be selected can be determined based on the characteristics of the target battery and the availability of data.

[0047] After the initial model is selected, the next step is to optimize the initial battery mechanism model based on the state data corresponding to the target battery under multiple working conditions. The optimization process involves adjusting the parameters in the model so that the model's prediction results are as close as possible to the actual collected state data. Parameter adjustment can be performed by a variety of methods, including but not limited to least squares, genetic algorithms, particle swarm optimization algorithms, etc. The goal of optimization is usually to reduce the error between the model prediction value and the actual measurement value, so that the model can more accurately reflect the behavior of the battery under different working conditions.

[0048] As an optional embodiment, when the target battery mechanism model is an equivalent circuit model, the initial state of charge value of the target battery is determined based on the state data and the target battery mechanism model, including: determining the state equation of the equivalent circuit model based on Kirchhoff's voltage law, wherein the state equation characterizes the state of the equivalent circuit model, and the state equation includes the state of charge; based on the Kalman filter algorithm, predicting and updating the state equation multiple times until a preset stop condition is reached to obtain the target state equation; and determining the initial state of charge value of the target battery based on the target state equation.

[0049] Optionally, in this embodiment, a second-order equivalent circuit model can be selected as the target battery mechanism model. The equivalent circuit model is a circuit composed of an RC environment, an ohmic resistor, an ideal voltage source, etc. The RC environment is used to describe the dynamic characteristics of the battery, and the ohmic internal resistance and the ideal voltage source describe the static characteristics of the battery. Therefore, the equivalent circuit model has good model accuracy and can meet the accuracy and computational complexity required for real-time SOC estimation. Figure 4 is a circuit diagram of a second-order equivalent circuit model provided according to an optional embodiment of the present invention, such as Figure 4As shown, the typical second-order equivalent circuit model is based on two RC links and ohmic internal resistance. One RC model has a shorter time constant and is used to describe the electrochemical polarization process, while the other RC model has a larger time constant and is used to describe concentration polarization to meet the needs of the experiment.

[0050] Based on Kirchhoff's voltage law, the mathematical model of the equivalent circuit is expressed as follows:

[0051]

[0052] The above equation is discretized using the Laplace discrete formula to obtain the state space model of the second-order equivalent circuit model as follows:

[0053]

[0054] This equation has good steady-state and dynamic performance. The system observation equation is as follows:

[0055] U L,k =U ocv,k -U p1,k -U p2,k -U o,k +v

[0056] Among them, U L,k is the estimated value of the terminal voltage of the Kalman filter model at time t; U p1,k and U p2,k is the estimated voltage value of electrochemical polarization internal resistance and concentration polarization internal resistance:

[0057]

[0058] X (i) is the state vector of the i-th Sigma point at time k; n is the number of state vectors; λ is the scaling parameter used to reduce the total prediction error; α and β are the distribution states of the selected control sampling points; P(k|k) is the covariance matrix of the state variables of the Sigma point set at time k:

[0059]

[0060] Among them, X (i) (kk) is the i-th point in the Sigma point set at time k before one-step prediction;

[0061] X (i) (k+1k)=F[X (i) (kk),u(k)]

[0062] Among them, X (i) (k+1k) is the further prediction of the state vector of the forward k+1Sigma point at time k; F[] is the recursive state equation;

[0063]

[0064] P(k+1k) is the one-step prediction and covariance matrix, ω (i) Predict the weight size for the Sigma point set.

[0065] Z (i) (k+1,k)=h[X (i) (k+1,k)]

[0066]

[0067]

[0068] Among them, Z (i) (k+1,k) is X (i) (k+1|k) One-step prediction of the observation value of the i-th Sigma point of the state vector, h[] is the recursive observation equation; is the covariance matrix between observations; is the covariance matrix of observation values ​​and state values.

[0069]

[0070] P(k+1|k+1)=P(k+1|k)-K(k+1)P zkzk K T (k+1)

[0071] Wherein, K(k+1) is the Kalman filter gain matrix. According to the above-mentioned Kalman filter based on the second-order equivalent circuit model, the initial SOC value based on the mechanism model can be calculated.

[0072] As an optional embodiment, before the initial state of charge value is input into the target neural network model, it includes: obtaining multiple initial state of charge values ​​determined based on the target battery mechanism model and multiple real state of charge values ​​obtained based on battery testing equipment; inputting the multiple initial state of charge values ​​and the multiple real state of charge values ​​as sample sets into the initial neural network model for training to obtain the target neural network model.

[0073] Optionally, the training target neural network model can take the lithium battery state of charge data estimated by the above-mentioned battery mechanism-based model as input, and the actual state data obtained by the battery testing equipment as the target, and establish and train the neural network algorithm by converting the problem into a regression problem, obtain the correction value of the estimation result, and establish a mapping function between the predicted value of the battery mechanism model and the actual value.

[0074] Figure 5is a flow chart of initial SOC value result correction of a mechanism model based on a long short-term memory neural network according to an optional embodiment of the present invention, such as Figure 5 As shown, the SOC estimation value of the Kalman filter algorithm based on the equivalent circuit and the real SOC value obtained from the battery test equipment constitute a sample set (SOC_estimate, SOC_real), and the mapping function of SOC_estimate→SOC_real is found. First, the sample set is divided into a training set and a test set, and the model is trained with the training set data. The training steps include setting the initial settings of the model, setting the initial parameters, model training, updating the weights and biases, and obtaining the optimized model. Then, the test set data is corrected by the obtained optimized model. Finally, the accuracy analysis of the correction result is performed to obtain the mapping function of SOC_estimate→SOC_real.

[0075] Among them, a recurrent neural network based on LSTM (Long Short-Term Memory) can be constructed, and the structure of the neural network includes a sequence input layer, an LSTM layer, a Dense layer, and a sequence output layer. During training, the loss function selects the MAE (Mean Absolute Error) function, the training batch is set to 20, the evaluation function selects MAE and RMSE (Root Mean Squared Error), the optimizer selects the Adam algorithm, the number of iterations is set to 100, and the neural network is trained with the training set. Figure 6 is a curve diagram of the change of the loss function during the training process provided by an optional embodiment of the present invention, such as Figure 6 As shown, after a finite number of iterations, the network's loss function converges rapidly, and the mapping function y=f(x) from the estimated value of the mechanism model to the true value is obtained. The test set is used to verify the training effect of the neural network.

[0076] As an optional embodiment, the initial neural network model includes one of the following: Elman neural network, long short-term memory neural network, recursive neural network.

[0077] Optionally, the initial neural network model can select a neural network with a memory link, including but not limited to Elman neural network, LSTM recurrent neural network, and RNN recursive neural network. Among them, Elman neural network is a recursive neural network with time memory function. It adds memory units, namely context units, to the traditional feedforward neural network to store the network state of the previous moment, so that it can process input data with time series characteristics; LSTM is a special form of RNN, which effectively solves the gradient vanishing or gradient exploding problem of traditional RNN when processing long sequence data by introducing a "gating" mechanism (such as input gate, forget gate and output gate); RNN is a neural network model that can process sequence data. It uses the output of the previous moment as the input of the current moment, so that it can capture the time dependency of the data. The three neural network models have their own advantages. In the SOC estimation of lithium-ion batteries, the selection of a suitable model depends on the complexity of battery behavior, the characteristics of the data, and the real-time computing requirements of the model. For example, in the case of long-term memory and processing of complex sequence data, LSTM may be a better choice; when the data sequence is short or the real-time requirements are high, Elman neural network or recursive neural network may be more applicable.

[0078] As an optional embodiment, a method for estimating the state of charge of a lithium-ion battery based on a combination of a mechanism model and a neural network is also provided. The method of the battery mechanism model is first used for estimation, which is then used as the input value of the neural network. The corresponding state of the lithium-ion battery obtained from actual testing is used as the target value. A training set is established using a random sampling method, and a neural network model with a time memory link is established and trained to obtain a mapping function between the predicted value and the true value, thereby optimizing the battery mechanism model's estimation effect on the state of charge, and further achieving accurate evaluation of the state of charge of the lithium-ion battery. Figure 7 is a flow chart of a method for estimating the state of charge of a lithium-ion battery based on a combination of a mechanism model and a neural network according to an optional embodiment of the present invention, such as Figure 7 As shown, the specific steps are as follows:

[0079] The steps for acquiring dynamic operating condition data of lithium-ion batteries refer to conducting charge and discharge tests and pulse charge and discharge tests under dynamic operating conditions of lithium batteries based on professional battery testing equipment to obtain the real status data of lithium batteries.

[0080] The step of estimating the state of charge of a lithium-ion battery based on a battery mechanism model refers to establishing a battery mechanism model and using the battery dynamic operating condition charge and discharge and pulse charge and discharge test data to estimate the state of charge of the lithium battery under the current operating condition.

[0081] The neural network model construction and training step refers to taking the lithium battery state of charge data estimated based on the battery mechanism model as input, taking the lithium battery state of charge data obtained by the battery testing equipment as the target, taking the lithium battery state of charge estimation data and the lithium battery state of charge data as sample sets and dividing them into training set data and test set data, and establishing and training the neural network model through the training set data and the test set data, that is, establishing a mapping function between the estimated value of the lithium battery state of charge data in the battery mechanism model and the true value of the state of charge data obtained by the battery testing equipment.

[0082] The step of constructing a state parameter estimation algorithm that integrates mechanism and data refers to estimating the actual battery state data under dynamic working conditions of the lithium battery collected by the actual vehicle sensor by first using the battery mechanism model in the lithium-ion battery state of charge estimation step based on the battery mechanism model to estimate the lithium battery state of charge parameters, and then correcting the estimation result with the neural network algorithm in the neural network model construction and training step to obtain the final estimated value of the lithium-ion battery state of charge.

[0083] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0084] Through the description of the above implementation methods, those skilled in the art can clearly understand that the charge state determination method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0085] According to an embodiment of the present invention, there is also provided a vehicle-mounted embedded system, which applies any one of the above-mentioned state of charge determination methods, including: a mechanism model module, used to determine the initial state of charge value of the target battery; and an offline network module, used to correct the initial state of charge value based on the target neural network model to obtain the target state of charge value of the target battery.

[0086] Optionally, the battery mechanism model and the data-driven neural network model can be coupled, and the trained neural network algorithm can be applied offline to the vehicle embedded system. Among them, the mechanism model module is the first-level SOC estimation unit in the system, which predicts the SOC based on the physical and chemical properties of the battery. In this module, a second-order equivalent circuit model can be used, which includes a voltage source, an ohmic internal resistance, and two RC links, which can describe the static and dynamic characteristics of the battery; the offline network module is a supplement to the mechanism model module. It corrects the initial SOC estimation value of the mechanism model module based on the neural network to improve the accuracy of the SOC estimation. The vehicle embedded system integrates the mechanism model module and the offline network module to realize a dual SOC estimation mechanism. The mechanism model module can provide SOC estimation based on physical principles, while the offline network module corrects the estimation error of the mechanism model through a data-driven method, thereby improving the accuracy of the SOC estimation.

[0087] According to an embodiment of the present invention, a state of charge determination device for implementing the above state of charge determination method is also provided. Figure 8 is a structural block diagram of a device for determining a state of charge according to an embodiment of the present invention. Figure 8 As shown, the charge state determination device includes: an acquisition module 81, a construction module 82, a determination module 83 and a correction module 84. The charge state determination device is described below.

[0088] The acquisition module 81 is used to acquire the status data of the target battery, wherein the status data includes the voltage, current and internal resistance of the target battery.

[0089] The construction module 82 is connected to the acquisition module 81 and is used to construct a target battery mechanism model corresponding to the target battery.

[0090] The determination module 83 is connected to the construction module 82 and is used to determine the initial state of charge value of the target battery based on the state data and the target battery mechanism model.

[0091] The correction module 84 is connected to the determination module 83 and is used to input the initial state of charge value into the target neural network model to obtain the target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0092] It should be noted that the acquisition module 81, construction module 82, determination module 83 and correction module 84 correspond to steps S201 to S204 in the embodiment, and the examples and application scenarios implemented by the multiple modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules, as part of the device, can be run in the computer terminal 10 provided in the embodiment.

[0093] An embodiment of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one network device among multiple network devices of a computer network. The computer device includes a memory and a processor.

[0094] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the state of charge determination method and device in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, the above-mentioned state of charge determination method is realized. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories can be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0095] The processor can call the information and application programs stored in the memory through the transmission device to perform the following steps: obtain the status data of the target battery, wherein the status data includes the voltage, current and internal resistance of the target battery; construct a target battery mechanism model corresponding to the target battery; determine the initial state of charge value of the target battery based on the status data and the target battery mechanism model; input the initial state of charge value into the target neural network model to obtain the target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0096] Optionally, the processor may also execute the program code of the following steps: constructing a target battery mechanism model corresponding to the target battery, including: obtaining state data corresponding to the target battery under multiple working conditions; based on the target battery, selecting an initial battery mechanism model, wherein the initial battery mechanism model includes one of the following: an equivalent circuit model, an electrochemical model, and an empirical formula model; based on the state data corresponding to the target battery under multiple working conditions, optimizing the initial battery mechanism model to obtain the target battery mechanism model.

[0097] Optionally, the processor may also execute the program code of the following steps: when the target battery mechanism model is an equivalent circuit model, determining the initial state of charge value of the target battery based on the state data and the target battery mechanism model, including: determining the state equation of the equivalent circuit model based on Kirchhoff's voltage law, wherein the state equation characterizes the state of the equivalent circuit model, and the state equation includes the state of charge; based on the Kalman filter algorithm, predicting and updating the state equation multiple times until a preset stop condition is reached to obtain the target state equation; determining the initial state of charge value of the target battery based on the target state equation.

[0098] Optionally, the processor may also execute program code for the following steps: before inputting the initial state of charge value into the target neural network model, it includes: obtaining multiple initial state of charge values ​​determined based on the target battery mechanism model and multiple real state of charge values ​​obtained based on battery testing equipment; inputting the multiple initial state of charge values ​​and the multiple real state of charge values ​​as sample sets into the initial neural network model for training to obtain the target neural network model.

[0099] Optionally, the processor may also execute program code of the following steps: the initial neural network model includes one of the following: an Elman neural network, a long short-term memory neural network, and a recursive neural network.

[0100] According to an embodiment of the present invention, a method for determining the state of charge is provided. The state data of a target battery is obtained, wherein the state data includes the voltage, current and internal resistance of the target battery; a target battery mechanism model corresponding to the target battery is constructed; based on the state data and the target battery mechanism model, an initial state of charge value of the target battery is determined; the initial state of charge value is input into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value, thereby achieving the purpose of accurately estimating the state of charge of the battery, thereby achieving the technical effect of improving the accuracy and efficiency of the state of charge estimation, and further solving the technical problem of large errors in estimating the state of charge using the battery mechanism model.

[0101] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a non-volatile storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0102] The embodiment of the present invention further provides a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium can be used to store the program code executed by the charge state determination method provided in the above embodiment.

[0103] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0104] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: obtaining status data of a target battery, wherein the status data includes voltage, current, and internal resistance of the target battery; constructing a target battery mechanism model corresponding to the target battery; determining an initial state of charge value of the target battery based on the status data and the target battery mechanism model; inputting the initial state of charge value into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0105] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: constructing a target battery mechanism model corresponding to the target battery, including: obtaining state data corresponding to the target battery under multiple working conditions; based on the target battery, selecting an initial battery mechanism model, wherein the initial battery mechanism model includes one of the following: an equivalent circuit model, an electrochemical model, and an empirical formula model; based on the state data corresponding to the target battery under multiple working conditions, optimizing the initial battery mechanism model to obtain the target battery mechanism model.

[0106] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: when the target battery mechanism model is an equivalent circuit model, determining the initial state of charge value of the target battery based on the state data and the target battery mechanism model, including: determining the state equation of the equivalent circuit model based on Kirchhoff's voltage law, wherein the state equation characterizes the state of the equivalent circuit model, and the state equation includes the state of charge; based on the Kalman filtering algorithm, predicting and updating the state equation multiple times until a preset stop condition is reached to obtain the target state equation; determining the initial state of charge value of the target battery based on the target state equation.

[0107] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: before inputting the initial state of charge value into the target neural network model, including: obtaining multiple initial state of charge values ​​determined based on a target battery mechanism model and multiple real state of charge values ​​obtained based on a battery testing device; inputting the multiple initial state of charge values ​​and the multiple real state of charge values ​​as sample sets into the initial neural network model for training to obtain a target neural network model.

[0108] Optionally, in this embodiment, the non-volatile storage medium is configured to store program codes for executing the following steps: the initial neural network model includes one of the following: an Elman neural network, a long short-term memory neural network, a recursive neural network.

[0109] An embodiment of the present invention also provides a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it can achieve: obtaining status data of a target battery, wherein the status data includes voltage, current and internal resistance of the target battery; constructing a target battery mechanism model corresponding to the target battery; determining an initial state of charge value of the target battery based on the status data and the target battery mechanism model; inputting the initial state of charge value into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

[0110] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0111] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0112] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0113] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0114] In addition, each functional unit in each embodiment of the present invention 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. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0115] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or 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, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0116] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for determining a state of charge, characterized in that: include: Acquiring status data of a target battery, wherein the status data includes voltage, current, and internal resistance of the target battery; Constructing a target battery mechanism model corresponding to the target battery; Determining an initial state of charge value of the target battery based on the state data and the target battery mechanism model; The initial state of charge value is input into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

2. The method according to claim 1, characterized in that The constructing a target battery mechanism model corresponding to the target battery includes: Acquire state data corresponding to each of the target batteries under multiple working conditions; Based on the target battery, selecting an initial battery mechanism model, wherein the initial battery mechanism model includes one of the following: an equivalent circuit model, an electrochemical model, and an empirical formula model; Based on the state data corresponding to the target battery under multiple working conditions, the initial battery mechanism model is optimized to obtain the target battery mechanism model.

3. The method according to claim 1, characterized in that In the case where the target battery mechanism model is an equivalent circuit model, determining the initial state of charge value of the target battery based on the state data and the target battery mechanism model includes: Based on Kirchhoff's voltage law, determining a state equation of the equivalent circuit model, wherein the state equation represents a state of the equivalent circuit model, and the state equation includes a state of charge; Based on the Kalman filter algorithm, the state equation is predicted and updated multiple times until a preset stop condition is reached to obtain a target state equation; Based on the target state equation, an initial state of charge value of the target battery is determined.

4. The method according to claim 1, characterized in that: Before inputting the initial state of charge value into the target neural network model, the method includes: Acquire a plurality of initial state-of-charge values ​​determined based on the target battery mechanism model and a plurality of actual state-of-charge values ​​obtained based on a battery test device; A plurality of initial state of charge values ​​and a plurality of true state of charge values ​​are input as sample sets into an initial neural network model for training to obtain the target neural network model.

5. The method according to claim 4, characterized in that The initial neural network model includes one of the following: Elman neural network, long short-term memory neural network, recursive neural network.

6. A vehicle-mounted embedded system, characterized in that: The method for determining the state of charge according to any one of claims 1 to 5 comprises: A mechanism model module is used to determine the initial state of charge value of the target battery; The offline network module is used to correct the initial state of charge value based on the target neural network model to obtain the target state of charge value of the target battery.

7. A device for determining a state of charge, characterized in that: include: An acquisition module, used to acquire status data of a target battery, wherein the status data includes voltage, current and internal resistance of the target battery; A construction module, used to construct a target battery mechanism model corresponding to the target battery; A determination module, configured to determine an initial state of charge value of the target battery based on the state data and the target battery mechanism model; A correction module is used to input the initial state of charge value into a target neural network model to obtain a target state of charge value of the target battery, wherein the target neural network model is used to correct the initial state of charge value.

8. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute the charge state determination method according to any one of claims 1 to 5.

9. A computer device, characterized in that: include: Memory and processor, The memory stores a computer program; The processor is used to execute a computer program stored in the memory, and when the computer program is run, the processor executes the charge state determination method according to any one of claims 1 to 5.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for determining the state of charge according to any one of claims 1 to 5 is implemented.

Images