A method and device for identifying faults of a battery pack under random working conditions

Abstract

This invention relates to a method and apparatus for fault identification of battery packs under random operating conditions. The method includes: constructing a battery pack state model and a battery cell deviation model; using filters to predict the state of the battery pack state model and the battery cell deviation model respectively, obtaining a first prediction result for the battery pack at the next time step and a first prediction result for the battery cells at the next time step; inputting the obtained battery state parameters and vehicle state parameters into a trained time-series network prediction model to obtain a second prediction result for the battery pack at the next time step; inputting the second prediction result for the battery pack and the first prediction result for the battery cells into a trained physical augmentation encoder to obtain a second prediction matrix for the battery cells; calculating the residual value of the second prediction matrix for the battery cells and comparing it with a preset threshold to identify faults in the battery pack. This solution can achieve accurate identification of battery pack faults under random operating conditions.

Description

Technical Field

[0001] This invention relates to the field of power battery safety monitoring technology, and in particular to a method and apparatus for fault identification of battery packs under random operating conditions. Background Technology

[0002] Power battery systems composed of lithium-ion battery packs have advantages such as high specific energy, long lifespan, low self-discharge, and wide operating temperature range, making them the mainstream choice for electric vehicles. However, as the market share of electric vehicles increases, battery module safety failures have become a major bottleneck restricting their development. In recent years, frequent battery failures have led to countless recalls, increasing user safety anxiety and the economic burden on the market.

[0003] Under normal circumstances, with the increase of charge-discharge cycles, lithium-ion batteries experience irreversible losses in their microscopic electrochemical components, resulting in increased internal resistance and decreased capacity. However, due to the influence of random operating conditions and abuse tolerance, there is a certain probability of abnormal faults such as leakage, abnormal aging, or even extreme safety accidents such as battery short circuits, fires, and explosions during the battery's charge-discharge process. Abnormal battery degradation is a process that progresses from microscopic reaction abnormalities to macroscopic characterization abnormalities, and finally to an uncontrolled system. The uncontrollable causes of faults under actual operating conditions lead to significant differences in the duration of battery degradation and a highly random nature of critical trigger points. Furthermore, the types and quality of monitorable data during battery application are very limited, which undoubtedly greatly increases the difficulty of fault diagnosis.

[0004] In related technologies, most lithium-ion battery fault identification methods are conducted under experimental simulation conditions. However, in actual battery state monitoring, it is necessary to consider the influence of various factors such as the number of sensors, data quality, user behavior, and random operating conditions on the battery state. This makes it impossible to apply the diagnostic methods in laboratory research. Furthermore, existing battery fault diagnosis algorithms are difficult to balance the accuracy and transferability of fault identification.

[0005] Therefore, there is an urgent need to provide a method and device for fault identification of battery packs under random operating conditions. Summary of the Invention

[0006] To address the problem that traditional lithium-ion battery fault identification methods struggle to accurately identify battery pack faults under various random operating conditions, this invention provides a method and apparatus for identifying battery pack faults under random operating conditions.

[0007] In a first aspect, embodiments of the present invention provide a fault identification method for a battery pack under random operating conditions, the method comprising:

[0008] For each current moment, execute:

[0009] The voltage parameters of individual battery cells obtained at the current moment are optimized to obtain the battery pack voltage and individual battery cell deviation voltage at the current moment.

[0010] Based on the battery pack voltage and the individual cell deviation voltage, a battery pack state model and an individual cell deviation model are constructed.

[0011] The state prediction of the battery pack state model and the battery cell deviation model are performed using filters to obtain the first prediction result of the battery pack at the next time step and the first prediction result of the battery cell at the next time step, respectively.

[0012] The battery state parameters and vehicle state parameters obtained at the current moment are input into the trained temporal network prediction model to obtain the second prediction result of the battery pack at the next moment; wherein, during the training process of the temporal network prediction model, the first prediction result of the battery pack is used as a loss function to adjust the second prediction result in reverse;

[0013] The second prediction result of the battery pack and the first prediction result of the battery cell are input into the trained physical augmentation encoder to obtain the second prediction matrix of the battery cell;

[0014] The residual value of the second prediction matrix of the battery cell is calculated and compared with a preset threshold to identify faults in the battery pack.

[0015] Secondly, embodiments of the present invention also provide a fault identification device for a battery pack under random operating conditions, the device comprising:

[0016] The optimization unit is used to perform distribution optimization processing on the voltage parameters of the battery cells obtained at the current moment to obtain the battery pack voltage and the deviation voltage of the battery cells at the current moment.

[0017] The model building unit is used to build a battery pack state model and a battery cell deviation model based on the battery pack voltage and the battery cell deviation voltage.

[0018] The first prediction unit is used to perform state prediction on the battery pack state model and the battery cell deviation model using filters, respectively, to obtain the first prediction result of the battery pack at the next time and the first prediction result of the battery cell at the next time.

[0019] The second prediction unit is used to input the battery state parameters and vehicle state parameters obtained at the current moment into the trained temporal network prediction model to obtain the second prediction result of the battery pack at the next moment; wherein, during the training process of the temporal network prediction model, the first prediction result of the battery pack is used as a loss function to adjust the second prediction result in reverse.

[0020] The third prediction unit is used to input the second prediction result of the battery pack and the first prediction result of the battery cell into the trained physical enhancement encoder to obtain the second prediction matrix of the battery cell.

[0021] The fault identification unit is used to calculate the residual value of the second prediction matrix of the battery cell and compare it with a preset threshold to identify faults in the battery pack.

[0022] Thirdly, embodiments of the present invention also provide a computing device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, it implements the method described in any embodiment of this specification.

[0023] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the methods described in any embodiment of this specification.

[0024] Fifthly, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the steps of the method described above.

[0025] This invention provides a method and apparatus for fault identification of battery packs under random operating conditions. First, the voltage parameters of individual battery cells obtained from a cloud data platform are distributed and optimized. Based on the optimized voltage data of the battery pack and individual cells, a battery pack state model and an individual cell deviation model are constructed, respectively. Then, filters are used to predict the states of the battery pack and individual cells, obtaining prediction results with good stability. Next, a pre-trained time-series network prediction model is used to perform a secondary prediction of the battery pack state. The prediction results of the time-series network prediction model are inversely adjusted using the prediction results of the filters during training. Finally, a pre-trained physical augmentation encoder is used to further compress and decompress the prediction results of the filters and the time-series network model, thereby obtaining individual cell prediction results with both good stability and accuracy. Thus, this invention sequentially uses filters, a time-series network prediction model, and a physical augmentation encoder to predict the state of the battery pack, using massive amounts of cloud data as a training sample set, thereby ensuring that the prediction results balance robustness and accuracy, and ultimately achieving accurate identification of fault states in battery packs under random operating conditions. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0027] Figure 1 This is a flowchart of a fault identification method for a battery pack under random operating conditions provided by an embodiment of the present invention;

[0028] Figure 2 This is a hardware architecture diagram of a computing device provided in an embodiment of the present invention;

[0029] Figure 3 This is a structural diagram of a fault identification device for a battery pack under random operating conditions provided in an embodiment of the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0031] The following describes the specific implementation of the above concept.

[0032] Please refer to Figure 1 This invention provides a method for fault identification of battery packs under random operating conditions, the method comprising:

[0033] For each current moment, execute:

[0034] Step 100: Perform distribution optimization processing on the voltage parameters of the battery cells obtained at the current moment to obtain the battery pack voltage and battery cell deviation voltage at the current moment.

[0035] Step 102: Based on the battery pack voltage and the individual cell deviation voltage, construct the battery pack state model and the individual cell deviation model;

[0036] Step 104: Use filters to perform state prediction on the battery pack state model and the battery cell deviation model respectively, and obtain the first prediction result of the battery pack at the next time and the first prediction result of the battery cell at the next time; wherein, the first prediction result includes the voltage and state of charge of the battery pack and the battery cell at the next time.

[0037] Step 106: Input the battery state parameters and vehicle state parameters obtained at the current moment into the trained temporal network prediction model to obtain the second prediction result of the battery pack at the next moment; wherein, during the training process of the temporal network prediction model, the first prediction result of the battery pack is used as the loss function to adjust the second prediction result in reverse;

[0038] Step 108: Input the second prediction result of the battery pack and the first prediction result of the battery cell into the trained physical augmentation encoder to obtain the second prediction matrix of the battery cell;

[0039] Step 110: Calculate the residual value of the second prediction matrix of the battery cell and compare it with a preset threshold to identify faults in the battery pack.

[0040] In this embodiment of the invention, the voltage parameters of individual battery cells obtained from a cloud data platform are first subjected to distribution optimization processing. Based on the optimized voltage data of the battery pack and individual battery cells, a battery pack state model and an individual battery cell deviation model are constructed respectively. Then, filters are used to predict the state of the battery pack and individual battery cells respectively, thereby obtaining prediction results with good stability. Next, a pre-trained time-series network prediction model is used to predict the state of the battery pack a second time. In the training of the time-series network prediction model, the prediction results of the filter are used to adjust the prediction results of the time-series network model in reverse, thereby enhancing the accuracy of the prediction results. Finally, a pre-trained physical augmentation encoder is used to further compress and decompress the prediction results of the filter and the prediction results of the time-series network model, thereby obtaining a battery cell prediction result with both good stability and accuracy. Thus, this embodiment of the invention sequentially uses filters, time-series network prediction models, and physical augmentation encoders to predict the state of the battery pack, and uses massive cloud data as a training sample set, thereby making the prediction results both robust and accurate, and thus realizing the accurate identification of the fault state of the battery pack under random operating conditions.

[0041] For step 100:

[0042] In this embodiment of the invention, in order to analyze the overall state of the battery pack and each individual battery cell in the battery pack, it is necessary to perform distribution optimization processing on the voltage parameters of the individual battery cells obtained from the cloud data platform.

[0043] In some implementations, step 100 includes:

[0044] The inverse Gaussian distribution is used as the probability distribution fit for the voltage parameters of the battery cell, and the voltage distribution probability of the battery cell is calculated.

[0045] Based on the voltage distribution probability of the individual battery cells, the current voltage of the battery pack and the deviation voltage of the individual battery cells are obtained.

[0046] Due to differences in production batches of battery cells or the mutual influence between cells in a local circuit, the distribution of battery cell voltage often exhibits a skewed pattern. Therefore, in this embodiment of the invention, an inverse Gaussian distribution is selected as the probability distribution fitting for the battery cell voltage parameters. This allows for a detailed description of the distribution characteristics of battery cells without increasing the number of parameters to be solved.

[0047] In some preferred embodiments, the voltage distribution probability of a single battery cell is calculated using the following formula:

[0048]

[0049] In the formula, μ=U m λ=μ 3 / Var(U i ), U i U is the voltage of the battery cell. m The average voltage of the battery cell;

[0050] The voltage of the battery pack is corrected based on the obtained voltage distribution probability of the individual battery cells, thereby obtaining the battery pack voltage and the individual battery cell deviation voltage. In some preferred embodiments, the battery pack voltage and the individual battery cell deviation voltage are calculated using the following formulas:

[0051] U T，p =(f(U) i )*U i ) / ∑f(U i )

[0052] ΔU i =U i -U T，p

[0053] In the formula, U T，p The battery pack voltage, f(U) i ) represents the voltage distribution probability of a single battery cell, ΔU i The deviation voltage of the battery cell.

[0054] Regarding step 102:

[0055] In some implementations, the battery pack state model and the individual cell deviation state model are respectively as follows:

[0056]

[0057] ΔU i =ΔU OCV，i (ΔSOCi )-ΔR i I

[0058] In the formula, U T，p R is the terminal voltage of the battery pack state model. P，p C is the polarization internal resistance of the battery pack. P，p U is the polarization capacitor of the battery pack. P，p R is the polarization voltage of the battery pack. 0，p Let I be the ohmic resistance of the battery pack, and U be the instantaneous current. OCV，p ΔU is the open-circuit voltage of the battery pack. i Let ΔU be the terminal voltage of the single-cell deviation state model. OCV，i ΔSOC is the open-circuit voltage of the battery cell. i The state of charge of the battery cell is ΔR. i Ω represents the ohmic internal resistance of the battery cell.

[0059] Because the working circuit of the battery pack is extremely complex and the battery pack state is affected by many factors in actual vehicle operation, this embodiment uses the battery pack voltage and battery cell deviation obtained after distributed optimization as model output parameters to construct the above-mentioned battery pack and battery cell deviation state model. This model can serve as an equivalent model of complex battery circuits to dynamically describe the dynamic process of the battery pack system. Based on the above state model, a filter is further combined to achieve online closed-loop prediction of the battery pack and battery cell state.

[0060] Regarding step 104:

[0061] In some implementations, step 104 includes:

[0062] The coefficients in the battery pack state model and the battery cell deviation model at the current time are identified using the least multiplier method, so as to obtain the state parameters of the battery pack state model and the state parameters of the battery cell deviation model at the current time.

[0063] The state parameters of the battery pack state model and the state parameters of the battery cell deviation model at the current moment are respectively input into the filter to obtain the first prediction result of the battery pack at the next moment and the first prediction result of the battery cell at the next moment.

[0064] In this embodiment of the invention, it can be seen from the battery pack state model and the battery cell deviation state model that the state parameters of the battery pack state model include the polarization internal resistance, polarization capacitance, and ohmic resistance of the battery pack, while the state parameters of the battery cell deviation model include the ohmic internal resistance of the battery cell. In order to facilitate the identification of the state parameters of the battery pack state model and the battery cell deviation model, it is necessary to preprocess the battery pack state model and the battery cell deviation model.

[0065] Specifically, first define E T =U T，p -U OCV，p And by discretizing the battery pack state model, we obtain:

[0066] E T，p，k+1 =exp(-Δt / τ)E T，p，k +(-R O，p )I k+1 +(exp(-Δt / τ)R O，p -(1-exp(-Δt / τ))R P，p )I k

[0067] In the formula, Δt is the sampling interval of the i-th time, and τ = R P C P E is the time constant. T，p，k Let R be the equivalent electromotive force of the battery pack at time k. O，p I is the ohmic resistance of the battery pack. k+1 Let be the instantaneous current at time k+1;

[0068] Furthermore, define y p，k For the output matrix, Φ p，k Let θ be the input matrix. p，k If we consider the state parameter matrix, then the above equation can be further written as:

[0069]

[0070] In the formula, β1, β2, and β3 are all equivalent coefficients;

[0071] Similarly, after discretizing the above equation, we can obtain Δy. i，k =ΔΦ i，k Δθ i，k ;in,

[0072]

[0073] In the formula, Δy i，k , ΔΦ i，k , Δθ i，i All are equivalent matrices, ΔU OCV，i，kLet be the open-circuit voltage of the i-th battery cell at time k;

[0074] Based on this, the least-multiply-two algorithm can be used to identify the above coefficient matrix, thereby obtaining the state parameters of the battery pack state model and the state parameters of the battery cell deviation model. Then, the identified parameters are input into the filter (for example, a Kalman filter) to obtain the voltage and state of charge of the battery pack at the next time step, and the voltage and state of charge of the battery cell at the next time step.

[0075] Regarding step 106:

[0076] The method of predicting the state of the battery pack and individual cells by using a filter combined with a state model in step 104 has good stability, but the accuracy of its prediction results is relatively poor. Furthermore, in the state prediction results of the battery pack and individual cells obtained in step 104, the state of the battery pack has a greater impact on the entire battery pack. In this embodiment of the invention, a trained time-series network model is used to further predict the battery pack state prediction results output from the filter. This can enhance the accuracy of the battery pack system prediction, thereby making the prediction results both robust and accurate.

[0077] In some implementations, in step 106, the temporal network prediction model is trained in the following manner:

[0078] Construct a pre-defined temporal network learning model;

[0079] The hyperparameters and loss function of the temporal network learning model are set; wherein the loss function is as follows:

[0080]

[0081] In the formula, loss1 is the loss function, and t is the sampling time. The prediction results of the temporal network learning model. For the first prediction result of the battery pack, U T，p，k+1 Let x be the terminal voltage of the battery pack at time k+1. T，p，k+1 The state of charge of the battery pack at time k+1;

[0082] The known vehicle state parameters and battery state parameters are used as the input sample set, and the known battery pack voltage and state of charge parameters are used as the output sample set. The Adam algorithm is used to train the time-series network learning model. The vehicle state parameters include vehicle speed, vehicle mileage and vehicle operating status, and the battery state parameters include battery pack total voltage, current, battery pack average temperature and battery pack state of charge.

[0083] During training, the loss and accuracy of the temporal network learning model are calculated, and the network parameters of the temporal network learning model are iterated until the model converges or reaches the preset number of iterations, thus obtaining the temporal network prediction model.

[0084] In this embodiment of the invention, the known vehicle state parameters and battery state parameters are both sets of parameters collected by the cloud data platform under normal vehicle operating conditions. When training the time-series network learning model, massive cloud data is used as the training sample set, and the first prediction result of the battery pack output by the filter is used as the objective function to adjust the prediction result of the time-series network learning model in reverse. In this way, not only can battery pack state prediction be realized in random real-world scenarios, breaking through the limitation of randomness in data quality, but it also balances robustness and accuracy better than pure model judgment and data-driven algorithms. The battery pack prediction result output by the time-series network model has obvious advantages.

[0085] Regarding step 108:

[0086] In this embodiment of the invention, a physical enhancement encoder is first constructed and trained. This physical encoder is then used to compress and decompress the second prediction result of the battery pack and the first prediction result of the individual battery cells, and to predict the state of the individual battery cells. This allows for accurate identification of battery pack system faults using the predicted state of the individual battery cells.

[0087] In some preferred embodiments, the physical enhancement encoder includes an encoding module, a latent space sampling module, and a decoding module connected in sequence; wherein:

[0088] The encoding module is used to perform feature fusion and dimensionality reduction operations on the first prediction result of the input battery cell and the second prediction result of the battery pack to obtain latent variables;

[0089] The latent space sampling module is used to sum the latent variables output by the encoding module and the first prediction result of the battery cell to obtain sampling points corresponding to the number of battery cells.

[0090] The decoding module is used to map the sampling points corresponding to the number of battery cells output by the latent space sampling module to obtain the second prediction matrix of the battery cell.

[0091] In this embodiment of the invention, firstly, the coding module is designed to fuse the first prediction result of the battery cell output by the filter and the second prediction result of the battery pack output by the time-series network (hereinafter referred to as layer 1), thereby effectively avoiding the problems of low accuracy of the filter and poor robustness of the time-series network prediction model, thus achieving higher accuracy; at the same time, the current battery state data and vehicle state data are compressed (hereinafter referred to as layer 2) as a correction term in the latent space, playing a corrective role for weakly correlated physical quantities.

[0092] The calculation process of the encoding module is as follows:

[0093]

[0094] In the formula, For layer 1 output, For layer 1 weights, For Layer 1 input, For layer 1 bias, For layer 2 output, σ e For activation function, Input weights for layer 2, x k For Layer 2 input, This is for layer 2 bias.

[0095] Next, a latent space sampling module is designed to sum the latent variables output by the encoding module with the first prediction result of the battery cell obtained in step 104, thereby obtaining the sampling points corresponding to the number of battery cells (hereinafter described as layer 3). The calculation process is as follows:

[0096]

[0097] In the formula, For layer 3 output, The deviation state matrix, For layer 1 output, For layer 2 output, I T It is the identity matrix. The charge state deviation matrix, This is the voltage state deviation matrix;

[0098] Finally, the decoding module maps the sampling points output by the latent space sampling module to the data space of the future state of the battery pack system, thereby realizing the prediction of the state of individual battery cells (described below as layer 4). The designed decompression network is as follows:

[0099]

[0100] In the formula, The second prediction matrix of the battery cell output by the physical enhancement encoder, σ dThe activation function, designed to take into account battery operating conditions, is modified and added to a specific coefficient to ensure that the output range remains within a reasonable voltage representation range, effectively avoiding gradient explosion during training. For layer 4 weights, These are the sampling points output by the potential space sampling module, corresponding to the number of battery cells. It is biased for layer 4.

[0101] In some preferred embodiments, the physical augmentation encoder is trained using known vehicle state parameters and battery state parameters as input sample sets and known battery pack voltage and state of charge parameters as output sample sets.

[0102] It is understandable that both the physical augmentation encoder and the temporal network prediction model are trained on a training set consisting of normal vehicles, but they use different loss functions.

[0103] Specifically, during the training process of the physical augmentation encoder, the MSE function is used as a loss function to back-adjust the prediction results of the physical augmentation encoder; wherein, the loss function is:

[0104]

[0105] In the formula, loss2 is the MSE function. The second prediction result of the battery cell is the output of the physical enhancement encoder, where k is the sampling time. Here is the actual observation matrix of a single battery cell, x n，k For the nth battery cell, U n，k Let be the terminal voltage of the nth battery cell.

[0106] It should be noted that, in the embodiments of the present invention, the prediction results obtained by the filter, the time-series network prediction model and the physical enhancement encoder all include the voltage and state of charge of the battery pack and / or individual battery cells at the next moment.

[0107] Regarding step 110:

[0108] In this embodiment of the invention, a residual matrix is ​​obtained by subtracting the second prediction matrix of the battery cell output by the physical enhancement encoder from the actual observation matrix of the battery cell observed during vehicle operation. The residual matrix is ​​then calculated to obtain the residual value. The calculated residual value is compared with a preset threshold. If the residual value exceeds the preset threshold, it can be determined that there is an abnormal point in the battery pack, thereby realizing the identification of battery pack faults. At the same time, multiple preset thresholds can be set to further realize the quantification and classification of battery pack faults.

[0109] like Figure 2 , Figure 3 As shown, this embodiment of the invention provides a fault identification device for a battery pack under random operating conditions. The device embodiment can be implemented through software, hardware, or a combination of both. From a hardware perspective, as... Figure 2 The diagram shown is a hardware architecture diagram of a computing device housing a fault identification device for a battery pack under random operating conditions, as provided in an embodiment of the present invention. (Except for...) Figure 2 In addition to the processor, memory, network interface, and non-volatile memory shown, the computing device in the embodiment may also include other hardware, such as a forwarding chip responsible for processing packets. Taking software implementation as an example, such as... Figure 3 As shown, a device in a logical sense is formed by the CPU of its computing device reading the corresponding computer program from the non-volatile memory into memory and running it. This embodiment provides a fault identification device for a battery pack under random operating conditions, the device comprising:

[0110] The optimization unit 301 is used to perform distribution optimization processing on the voltage parameters of the battery cells obtained at the current moment to obtain the battery pack voltage and the battery cell deviation voltage at the current moment.

[0111] The model building unit 302 is used to build a battery pack state model and a battery cell deviation model based on the battery pack voltage and the battery cell deviation voltage.

[0112] The first prediction unit 303 is used to perform state prediction on the battery pack state model and the battery cell deviation model using filters, respectively, to obtain the first prediction result of the battery pack at the next time and the first prediction result of the battery cell at the next time; wherein, the first prediction result includes the voltage and state of charge of the battery pack and the battery cell at the next time.

[0113] The second prediction unit 304 is used to input the battery state parameters and vehicle state parameters obtained at the current moment into the trained temporal network prediction model to obtain the second prediction result of the battery pack at the next moment; wherein, during the training process of the temporal network prediction model, the first prediction result of the battery pack is used as a loss function to adjust the second prediction result in reverse;

[0114] The third prediction unit 305 is used to input the second prediction result of the battery pack and the first prediction result of the battery cell into the trained physical enhancement encoder to obtain the second prediction matrix of the battery cell.

[0115] The fault identification unit 306 is used to calculate the residual value of the second prediction matrix of the battery cell and compare it with a preset threshold to identify faults in the battery pack.

[0116] In one embodiment of the present invention, when the optimization unit 301 performs distribution optimization processing on the voltage parameters of the battery cell obtained at the current moment, it performs the following operations:

[0117] The inverse Gaussian distribution is used as the probability distribution of the battery parameters of the battery cell to calculate the voltage distribution probability of the battery cell.

[0118] Based on the voltage distribution probability of the individual battery cells, the current voltage of the battery pack and the deviation voltage of the individual battery cells are obtained.

[0119] In one embodiment of the present invention, the battery pack state model and the battery cell deviation state model in the model building unit 302 are as follows:

[0120]

[0121] ΔU i =ΔU OCV，i (ΔSOC i )-ΔR i I

[0122] In the formula, U T，p R is the terminal voltage of the battery pack state model. P，p C is the polarization internal resistance of the battery pack. P，p U is the polarization capacitor of the battery pack. P，p R is the polarization voltage of the battery pack. 0，p Let I be the ohmic resistance of the battery pack, and U be the instantaneous current. OCV，p ΔU is the open-circuit voltage of the battery pack. i Let ΔU be the terminal voltage of the single-cell deviation state model. OCV，i ΔSOC is the open-circuit voltage of the battery cell. i The state of charge of the battery cell is ΔR. i Ω represents the ohmic internal resistance of the battery cell.

[0123] In one embodiment of the present invention, when the first prediction unit 303 performs state prediction on the battery pack state model and the battery cell deviation model at the current time using filters, it performs the following operations:

[0124] The coefficients in the battery pack state model and the battery cell deviation model at the current time are identified using the least multiplier method, so as to obtain the state parameters of the battery pack state model and the state parameters of the battery cell deviation model at the current time.

[0125] The state parameters of the battery pack state model and the state parameters of the battery cell deviation model at the current moment are respectively input into the filter to obtain the first prediction result of the battery pack at the next moment and the first prediction result of the battery cell at the next moment.

[0126] In one embodiment of the present invention, the temporal network prediction model in the second prediction unit 304 is trained in the following manner:

[0127] Construct a pre-defined temporal network learning model;

[0128] The hyperparameters and loss function of the temporal network learning model are set; wherein the loss function is as follows:

[0129]

[0130] In the formula, loss1 is the loss function, and t is the sampling time. The prediction results of the temporal network learning model. For the first prediction result of the battery pack, U T，p，k+1 Let x be the terminal voltage of the battery pack at time k+1. T，p，k+1 The state of charge of the battery pack at time k+1;

[0131] The known vehicle state parameters and battery state parameters are used as the input sample set, and the known battery pack voltage and state of charge parameters are used as the output sample set. The Adam algorithm is used to train the time-series network learning model. The vehicle state parameters include vehicle speed, vehicle mileage and vehicle operating status, and the battery state parameters include battery pack total voltage, current, battery pack average temperature and battery pack state of charge.

[0132] During training, the loss and accuracy of the temporal network learning model are calculated, and the network parameters of the temporal network learning model are iterated until the model converges or reaches the preset number of iterations, thus obtaining the temporal network prediction model.

[0133] In one embodiment of the present invention, the third prediction unit 305 includes a physical enhancement encoder, a latent space sampling module, and a decoding module connected in sequence; wherein:

[0134] The encoding module is used to perform feature fusion and dimensionality reduction operations on the first prediction result of the input battery cell and the second prediction result of the battery pack to obtain latent variables;

[0135] The latent space sampling module is used to sum the latent variables output by the encoding module and the first prediction results of the battery cells to obtain sampling points corresponding to the number of battery cells.

[0136] The decoding module is used to map the sampling points output by the latent space sampling module corresponding to the number of battery cells to obtain the second prediction matrix of the battery cell.

[0137] In one embodiment of the present invention, in the third prediction unit 305, the physical enhancement encoder is trained using known vehicle state parameters and battery state parameters as input sample sets and known battery pack voltage and state of charge parameters as output sample sets.

[0138] During training, the MSE function is used as a loss function to back-adjust the prediction results of the physical augmentation encoder; wherein, the loss function is:

[0139]

[0140] In the formula, loss2 is the MSE function. The prediction result of the physical enhancement encoder is given, where k is the sampling time. Here is the actual observation matrix of a single battery cell, x n，k For the nth battery cell, U n，k Let be the terminal voltage of the nth battery cell.

[0141] It is understood that the structures illustrated in the embodiments of the present invention do not constitute a specific limitation on a fault identification device for a battery pack under random operating conditions. In other embodiments of the present invention, a fault identification device for a battery pack under random operating conditions may include more or fewer components than illustrated, or combine some components, or split some components, or have different component arrangements. The illustrated components may be implemented in hardware, software, or a combination of software and hardware.

[0142] The information interaction and execution process between the modules in the above-mentioned device are based on the same concept as the method embodiment of the present invention, and the specific details can be found in the description of the method embodiment of the present invention, and will not be repeated here.

[0143] This invention also provides a computing device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements a fault identification method for a battery pack under random operating conditions according to any embodiment of this invention.

[0144] This invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform a fault identification method for a battery pack under random operating conditions according to any embodiment of this invention.

[0145] Specifically, a system or apparatus equipped with a storage medium may be provided, on which software program code implementing the functions of any of the embodiments described above is stored, and the computer (or CPU or MPU) of the system or apparatus may read and execute the program code stored in the storage medium.

[0146] In this case, the program code read from the storage medium can itself implement the function of any of the above embodiments, and therefore the program code and the storage medium storing the program code constitute part of the present invention.

[0147] Examples of storage media used to provide program code include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Alternatively, program code can be downloaded from a server computer via a communication network.

[0148] Furthermore, it should be clear that not only can the program code read by the computer be executed, but also the operating system or other components operating on the computer can be instructed based on the program code to perform some or all of the actual operations, thereby realizing the function of any of the embodiments described above.

[0149] Furthermore, it is understood that the program code read from the storage medium is written to the memory set in the expansion board inserted into the computer or to the memory set in the expansion module connected to the computer. Then, based on the instructions of the program code, the CPU or other components installed on the expansion board or expansion module execute some and all of the actual operations, thereby realizing the function of any of the above embodiments.

[0150] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0151] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as ROM, RAM, magnetic disk, or optical disk.

[0152] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for fault identification of a battery pack under random operating conditions, characterized in that, include: For each current moment, execute: The voltage parameters of individual battery cells obtained at the current moment are optimized to obtain the battery pack voltage and individual battery cell deviation voltage at the current moment. Based on the battery pack voltage and the individual cell deviation voltage, a battery pack state model and an individual cell deviation model are constructed. The battery pack state model and the individual battery cell deviation state model are as follows: D.U. i =ΔU OCV，i (ΔSOC i )-ΔR i I In the formula, U T,p R is the terminal voltage of the battery pack state model. P，p C is the polarization internal resistance of the battery pack. P，p U is the polarization capacitor of the battery pack. P，p R is the polarization voltage of the battery pack. 0，p Let I be the ohmic resistance of the battery pack, and U be the instantaneous current. OCV，p ΔU is the open-circuit voltage of the battery pack. i Let ΔU be the terminal voltage of the single-cell deviation state model. OCV，i ΔSOC is the open-circuit voltage of the battery cell. i The state of charge of the battery cell is ΔR. i The ohmic internal resistance of the battery cell; The coefficients of the battery pack state model and the battery cell deviation model at the current time are identified using the least multiplier method, so as to obtain the state parameters of the battery pack state model and the state parameters of the battery cell deviation model at the current time. The state parameters of the battery pack state model and the state parameters of the battery cell deviation model at the current moment are respectively input into the Kalman filter, so as to use the Kalman filter to predict the state of the battery pack state model and the battery cell deviation model respectively, and obtain the first prediction result of the battery pack at the next moment and the first prediction result of the battery cell at the next moment respectively. The battery state parameters and vehicle state parameters obtained at the current moment are input into the trained temporal network prediction model to obtain the second prediction result of the battery pack at the next moment. During the training process of the temporal network prediction model, the first prediction result of the battery pack is used as a loss function to adjust the second prediction result in reverse. The vehicle state parameters include vehicle speed, vehicle mileage and vehicle operating status, and the battery state parameters include battery pack total voltage, current, battery pack average temperature and battery pack state of charge. The second prediction result of the battery pack and the first prediction result of the individual battery cells are input into the trained physical augmentation encoder to obtain the second prediction matrix of the individual battery cells; the physical augmentation encoder includes an encoding module, a latent space sampling module, and a decoding module connected in sequence; wherein: The encoding module is used to perform feature fusion and dimensionality reduction operations on the first prediction result of the input battery cell and the second prediction result of the battery pack to obtain latent variables; The latent space sampling module is used to sum the latent variables output by the encoding module and the first prediction results of the battery cells to obtain sampling points corresponding to the number of battery cells. The decoding module is used to map the sampling points output by the latent space sampling module corresponding to the number of battery cells to obtain the second prediction matrix of the battery cell. The residual value of the second prediction matrix of the battery cell is calculated and compared with a preset threshold to identify faults in the battery pack.

2. The method according to claim 1, characterized in that, The step of performing distribution optimization processing on the voltage parameters of individual battery cells obtained at the current moment to obtain the battery pack voltage and individual battery cell deviation voltage at the current moment includes: The inverse Gaussian distribution is used as the probability distribution of the battery parameters of the battery cell to calculate the voltage distribution probability of the battery cell. Based on the voltage distribution probability of the individual battery cells, the current voltage of the battery pack and the deviation voltage of the individual battery cells are obtained.

3. The method according to claim 1, characterized in that, The temporal network prediction model is trained in the following manner: Construct a pre-defined temporal network learning model; The hyperparameters and loss function of the temporal network learning model are set; wherein the loss function is as follows: In the formula, loss1 is the loss function, and t is the sampling time. The prediction results of the temporal network learning model. For the first prediction result of the battery pack, U T,p,k+1 Let x be the terminal voltage of the battery pack at time k+1. T,p,k+1 The state of charge of the battery pack at time k+1; The known vehicle state parameters and battery state parameters are used as the input sample set, and the known battery pack voltage and state of charge parameters are used as the output sample set. The Adam algorithm is used to train the time-series network learning model. The vehicle state parameters include vehicle speed, vehicle mileage and vehicle operating status, and the battery state parameters include battery pack total voltage, current, battery pack average temperature and battery pack state of charge. During training, the loss and accuracy of the temporal network learning model are calculated, and the network parameters of the temporal network learning model are iterated until the model converges or reaches the preset number of iterations, thus obtaining the temporal network prediction model.

4. The method according to claim 1, characterized in that, The physical augmentation encoder is trained using known vehicle state parameters and battery state parameters as input sample sets and known battery pack voltage and state of charge parameters as output sample sets. During training, the MSE function is used as a loss function to back-adjust the prediction results of the physical augmentation encoder; wherein, the loss function is: In the formula, loss2 is the MSE function. The prediction result of the physical enhancement encoder is given, where k is the sampling time. Here is the actual observation matrix of a single battery cell, x n,k For the nth battery cell, U n,k Let be the terminal voltage of the nth battery cell.

5. A fault detection device for a battery pack under random operating conditions, used to implement the method according to any one of claims 1 to 4, characterized in that, include: The optimization unit is used to perform distribution optimization processing on the voltage parameters of the battery cells obtained at the current moment to obtain the battery pack voltage and the deviation voltage of the battery cells at the current moment. The model building unit is used to build a battery pack state model and a battery cell deviation model based on the battery pack voltage and the battery cell deviation voltage. The first prediction unit is used to perform state prediction on the battery pack state model and the battery cell deviation model using filters, respectively, to obtain the first prediction result of the battery pack at the next time and the first prediction result of the battery cell at the next time. The second prediction unit is used to input the battery state parameters and vehicle state parameters obtained at the current moment into the trained temporal network prediction model to obtain the second prediction result of the battery pack at the next moment; wherein, during the training process of the temporal network prediction model, the first prediction result of the battery pack is used as a loss function to adjust the second prediction result in reverse. The third prediction unit is used to input the second prediction result of the battery pack and the first prediction result of the battery cell into the trained physical enhancement encoder to obtain the second prediction matrix of the battery cell. The fault identification unit is used to calculate the residual value of the second prediction matrix of the battery cell and compare it with a preset threshold to identify faults in the battery pack.

6. A computing device comprising a memory and a processor, wherein the memory stores a computer program, and the processor, when executing the computer program, implements the method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, causes the computer to perform the method of any one of claims 1-4.

Images