Digital twin lithium battery real-time state monitoring system

Abstract

The invention relates to the technical field of state monitoring, in particular to a real-time state monitoring system for a digital twin lithium battery, which comprises a multi-source heterogeneous data acquisition module configured with a distributed sensor array and outputting multi-physical field original data; the time-varying feature fusion module is used for receiving the multi-physical field original data and generating a fusion feature vector with a unified timestamp through a space-time registration algorithm; the dynamic twinborn modeling module is used for constructing a cross-scale digital twinborn model comprising an electrochemical-mechanical coupling effect; the three-dimensional state mapping module is used for outputting a three-dimensional state matrix comprising the negative expansion rate, the electrolyte decomposition degree and the interface reaction entropy change; the life degradation prediction module is used for calculating a battery remaining service life prediction value and failure probability distribution through a coupling degradation model; and the model self-optimization module is used for dynamically adjusting a parameter updating strategy of the digital twin model and outputting an optimized sensor sampling frequency instruction. The invention provides powerful support for prolonging the service life of the battery and guaranteeing the operation safety of the system.

Description

Technical Field

[0001] The present invention relates to the field of condition monitoring technology, and in particular to a digital twin lithium battery real-time condition monitoring system. Background Art

[0002] Existing lithium battery status monitoring technologies mainly rely on a single or a small number of sensors to obtain macro indicators such as voltage, current, and temperature. They lack real-time perception of internal microscopic change processes and find it difficult to accurately capture multi-physical field details such as changes in electrolyte viscosity, electrode strain evolution, and local temperature gradient fluctuations. Traditional monitoring methods generally use static modeling, which cannot effectively describe the electrochemical-mechanical-thermodynamic multi-scale coupling mechanism, resulting in a lag in the prediction of degradation phenomena such as SEI film growth, lithium ion migration, and stress accumulation, and the health status assessment results are prone to delays or distortion. In addition, existing methods mostly collect data at a fixed sampling frequency and lack the ability to dynamically optimize monitoring strategies based on the actual failure risk of the battery, which not only causes energy waste, but also limits the rapid response to sudden degradation events.

[0003] In terms of life prediction, most of the existing degradation models are based on empirical degradation curves or simplified assumptions, and fail to fully consider the combined effects of multiple factors such as negative electrode expansion, electrolyte decomposition, and interfacial reaction entropy change. The life estimation deviation is large, and the failure probability distribution prediction accuracy is insufficient. At the same time, the traditional model update method is fixed and passive, and it is impossible to adaptively adjust the twin modeling parameters according to the failure trend monitored in real time, which easily leads to long-term simulation error accumulation, making the prediction out of line with the actual life. Summary of the Invention

[0004] Based on the above objectives, the present invention provides a digital twin lithium battery real-time status monitoring system.

[0005] The digital twin lithium battery real-time status monitoring system includes the following modules:

[0006] Multi-source heterogeneous data acquisition module: Configured with a distributed sensor array, it outputs multi-physics field raw data including dynamic values ​​of electrolyte viscosity, electrode strain waveforms, and local temperature gradients;

[0007] Time-varying feature fusion module: receiving the multi-physics field raw data and generating a fusion feature vector with a unified timestamp through a spatiotemporal registration algorithm;

[0008] Dynamic twin modeling module: Based on the fused feature vector, a cross-scale digital twin model including electrochemical-mechanical coupling effects is constructed to output real-time parameters of lithium ion concentration field, SEI film three-dimensional morphology, and electrode stress field;

[0009] Three-dimensional state mapping module: converts the real-time parameters into battery health status indicators and outputs a three-dimensional state matrix including the negative electrode expansion rate, electrolyte decomposition degree, and interface reaction entropy change;

[0010] Life degradation prediction module: Based on the three-dimensional state matrix, the battery remaining service life prediction value and failure probability distribution are calculated through a coupled degradation model;

[0011] Model self-optimization module: dynamically adjusts the parameter update strategy of the digital twin model based on the failure probability distribution, and outputs the optimized sensor sampling frequency instruction to the data acquisition module.

[0012] Furthermore, the multi-source heterogeneous data acquisition module includes:

[0013] Distributed sensor deployment: sensor arrays are deployed in key locations of the lithium battery structure, including micro-electroviscosity sensor arrays in the electrolyte area, strain gauge sensors on the electrode surface, and MEMS temperature sensors in the gaps between battery cell layers.

[0014] Multi-physical quantity signal acquisition: Various sensors synchronously collect dynamic values ​​of electrolyte viscosity, electrode strain waveform, and local temperature gradient;

[0015] Time-scale unification of multi-source data: synchronize the timestamps and match the spatial labels of the collected data to construct a multi-physics field raw data set including viscosity, strain and temperature gradient information as input for the subsequent feature fusion module.

[0016] Furthermore, the unified time scale of multi-source data includes:

[0017] Timestamp synchronization processing: synchronize the sampling time of electrolyte viscosity value, electrode strain waveform, and local temperature gradient, and use dynamic time warping algorithm on the original sampling sequence to align them on the common time axis;

[0018] Spatial label matching: Map the sampling points to a unified spatial reference system based on the physical location of the sensor and set a unified spatial label set;

[0019] Construction of multi-physics field original data set: Construct a fusion data matrix of three types of physical quantities in a unified space-time coordinate system.

[0020] Furthermore, the time-varying feature fusion module includes:

[0021] Unified timestamp inheritance processing: Unified timestamp inheritance processing, time alignment of multi-source physical quantities, ensuring that subsequent processing is performed under the same time base to avoid repeated calculations;

[0022] Unified spatial registration and interpolation: The aligned physical quantity data are mapped to a unified spatial coordinate set, and the Gaussian kernel interpolation algorithm is used for spatial registration;

[0023] Fusion feature vector generation: The physical quantities after spatiotemporal registration are constructed into a fusion feature vector.

[0024] Furthermore, the dynamic twin modeling module includes:

[0025] Input feature-driven modeling initialization: Receive the fused feature vector output by the time-varying feature fusion module as the model input, drive the joint modeling process of electrochemical and mechanical behaviors, and establish the foundation for the coupling of the three physical fields of electricity, heat, and force;

[0026] Electrochemical-mechanical coupling modeling: Based on a set of coupled partial differential equations, it simulates the key internal physical processes of lithium batteries during operation.

[0027] Real-time output of state parameters: Based on the numerical solution of the partial differential coupled equations, three types of lithium battery state parameters are output at a set time step, including lithium ion concentration field, SEI film 3D morphology data, and electrode stress field.

[0028] Furthermore, the electrochemical-mechanical coupling modeling includes a lithium ion diffusion equation, a SEI film growth morphology modeling, and an electrode stress distribution model.

[0029] Furthermore, the three-dimensional state mapping module includes:

[0030] State parameter normalization processing: normalizing the lithium ion concentration field, SEI film thickness, electrode stress and temperature gradient to the interval [0,1];

[0031] Calculation of health status indicators: Calculate health status indicators based on normalized variables;

[0032] Three-dimensional state matrix construction and output: health status indicators are combined to form a three-dimensional state matrix.

[0033] Furthermore, the health status indicators include the negative electrode expansion rate, the electrolyte decomposition degree and the interface reaction entropy change.

[0034] Furthermore, the life degradation prediction module includes:

[0035] Health status matrix time series construction: Expand the three-dimensional health status matrix along the time dimension to construct the time series of each indicator;

[0036] Coupled degradation model calculation: Based on the multi-factor coupling characteristics of battery degradation, a weighted cumulative damage model is used to estimate the life consumption rate;

[0037] Remaining life and failure probability prediction: Calculate the remaining battery life and failure probability distribution based on the accumulated degradation.

[0038] Furthermore, the model self-optimization module includes:

[0039] Failure risk perception: Receives the failure probability distribution output by the life degradation prediction module and calculates the local failure growth rate within a set time window;

[0040] Dynamic parameter adjustment strategy generation: Dynamically adjust digital twin modeling parameters based on failure growth rate, using weighted adjustment coefficients to express the adjustment range;

[0041] Optimize sampling frequency instruction output: Synchronously adjust the sensor sampling frequency of the data acquisition module based on the failure risk level.

[0042] Beneficial effects of the present invention:

[0043] The present invention realizes real-time monitoring and precise modeling of the complex evolution process inside lithium batteries by constructing a full-process system of multi-source heterogeneous data collection, spatiotemporal feature fusion, dynamic twin modeling, health status mapping, life degradation prediction and model self-optimization. Compared with the existing monitoring method that only relies on a single physical quantity, the present invention can simultaneously capture the dynamic changes of multiple physical fields such as electrolyte viscosity, electrode strain and local temperature gradient, and combine cross-scale electrochemical-mechanical coupling modeling to output key internal parameters such as lithium ion concentration field, SEI film three-dimensional morphology and electrode stress distribution in real time, which greatly improves the timeliness and accuracy of lithium battery health status assessment, and effectively makes up for the problem that traditional monitoring methods are insufficient in identifying early degradation and microscopic damage evolution.

[0044] The present invention proposes a self-optimization mechanism that dynamically adjusts modeling parameters and sampling frequency based on the failure probability distribution, so that the monitoring system can automatically enhance or converge perception granularity according to the actual degradation risk, taking into account both monitoring accuracy and energy consumption control; by introducing multi-factor coupled degradation modeling and failure probability prediction, it can comprehensively consider the effects of negative electrode expansion rate, electrolyte decomposition degree and interface entropy change on life, and significantly improve the accuracy of remaining life prediction and the reliability of failure warning; overall, the present invention not only realizes dynamic health management of lithium batteries throughout their life cycle, but also provides strong support for extending battery service life and ensuring system operation safety, and has important engineering application value and industrial promotion prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0046] Figure 1 is a system module diagram of an embodiment of the present invention;

[0047] Figure 2 This is a data collection diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to specific embodiments.

[0049] It should be noted that, unless otherwise defined, the technical or scientific terms used in the present invention should have the usual meanings understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0050] like Figure 1-Figure 2 As shown in the figure, the digital twin lithium battery real-time status monitoring system includes the following modules:

[0051] Multi-source heterogeneous data acquisition module: Configured with a distributed sensor array, it outputs multi-physics field raw data including dynamic values ​​of electrolyte viscosity, electrode strain waveforms, and local temperature gradients;

[0052] Time-varying feature fusion module: receives raw data from multiple physical fields and generates fused feature vectors with a unified timestamp through a spatiotemporal registration algorithm;

[0053] Dynamic twin modeling module: Based on the fusion feature vector, it builds a cross-scale digital twin model including electrochemical-mechanical coupling effects, and outputs real-time parameters of lithium ion concentration field, SEI film 3D morphology, and electrode stress field;

[0054] 3D state mapping module: converts real-time parameters into battery health status indicators and outputs a 3D state matrix including negative electrode expansion rate, electrolyte decomposition degree, and interfacial reaction entropy change;

[0055] Life degradation prediction module: Based on the three-dimensional state matrix, the battery remaining service life prediction value and failure probability distribution are calculated through the coupled degradation model;

[0056] Model self-optimization module: Dynamically adjusts the parameter update strategy of the digital twin model based on the failure probability distribution, and outputs the optimized sensor sampling frequency instruction to the data acquisition module.

[0057] The multi-source heterogeneous data acquisition module includes:

[0058] Distributed sensor deployment: sensor arrays are deployed in key locations of the lithium battery structure, including micro-electroviscosity sensor arrays in the electrolyte area, strain gauge sensors on the electrode surface, and MEMS temperature sensors in the gaps between battery cell layers.

[0059] Multi-physical quantity signal acquisition: Various sensors simultaneously collect the dynamic value of electrolyte viscosity, electrode strain waveform, and local temperature gradient, including:

[0060] (1) The dynamic value of electrolyte viscosity μ(t) is output by the frequency response microviscosity sensor and is expressed as:

[0061]

[0062] Among them, F v (t) is the hysteresis resistance measured by the sensor (N), r is the radius of the detection microsphere (m), and v(t) is the instantaneous velocity of the microsphere (m / s);

[0063] (2) Electrode strain waveform ε(t), measured by the strain gauge, is expressed as:

[0064]

[0065] Where ΔL(t) is the change in electrode length per unit time (m), and L0 is the initial electrode length (m);

[0066] (3) Local temperature gradient The temperature difference calculated based on the multi-point MEMS temperature sensor is expressed as:

[0067]

[0068] Where T(x,t) is the temperature at position x (K), and Δx is the distance between adjacent sensors (m);

[0069] Time-scale unification of multi-source data: synchronize the timestamps and match the spatial labels of the collected data to construct a multi-physics field raw data set including viscosity, strain and temperature gradient information as input for the subsequent feature fusion module.

[0070] The unified time scale of multi-source data includes:

[0071] Timestamp synchronization processing: synchronize the sampling time of electrolyte viscosity value, electrode strain waveform, and local temperature gradient, and use linear interpolation or dynamic time warping (DTW) algorithm on the original sampling sequence to make it on the common time axis. The DTW alignment error calculation formula is:

[0072]

[0073] Among them, X i ,Y j represents the value of the two signals to be aligned at their respective moments, D(i,j) is the calculated error of the i-th and j-th points in the alignment cost matrix, and the final alignment path is the path with the minimum cumulative cost;

[0074] Spatial label matching: Based on the physical layout of the sensors (the spatial coordinates of the microviscosity sensor, electrode strain gauge, and MEMS temperature sensor), the sampling points are mapped to a unified spatial reference system and a unified spatial label set is set, which is expressed as:

[0075]

[0076] in, To unify the spatial label set, each data point is labeled (x i ,t j ), indicating that the physical quantity is at the spatial position x i , time point t j The state of the upper

[0077] Construction of multi-physics field original data set: Construct the fusion data matrix of the three types of physical quantities in a unified space-time coordinate system, expressed as:

[0078]

[0079] Among them, F(x i ,t j ) is the original data set of multi-physics field, μ(x i ,t j ) is the position x i , time t j The electrolyte viscosity value under i ,t j ) is the position x i , time t jThe electrode strain under is the position x i , time t j The local temperature gradient.

[0080] The time-varying feature fusion module includes:

[0081] Unified timestamp inheritance processing: Unified timestamp inheritance processing, time alignment of multi-source physical quantities, ensuring that subsequent processing is performed under the same time base to avoid repeated calculations;

[0082] Unified spatial registration and interpolation: Mapping the aligned physical data to a unified spatial coordinate set The Gaussian kernel interpolation algorithm is used for spatial registration, which is expressed as:

[0083]

[0084] Among them, F(x i ,t) is the original physical measurement value, F * (x, t) is the spatially aligned eigenvalue after interpolation, ω i (x) is the Gaussian kernel weighting factor, σ is the spatial smoothing factor, which represents the standard deviation of the Gaussian kernel function and reflects the degree of interpolation smoothness, and the unit is meter (m);

[0085] Fusion feature vector generation: The physical quantities after spatiotemporal registration are constructed into a fusion feature vector, which is expressed as:

[0086]

[0087] in, is the spatiotemporal fusion result of electrolyte viscosity, is the spatiotemporal fusion result of electrode strain, is the spatiotemporal fusion result of the local temperature gradient, and Z(x,t) is the fusion feature vector.

[0088] The dynamic twin modeling module includes:

[0089] Input feature-driven modeling initialization: Receive the fused feature vector output by the time-varying feature fusion module as the model input, drive the joint modeling process of electrochemical and mechanical behaviors, and establish the foundation for the coupling of the three physical fields of electricity, heat, and force;

[0090] Electrochemical-mechanical coupling modeling: Based on a set of coupled partial differential equations, it simulates the key internal physical processes of lithium batteries during operation.

[0091] Real-time output of state parameters: Based on the numerical solution of the partial differential coupled equations, three types of lithium battery state parameters are output at a set time step, including lithium ion concentration field c(x,t), SEI film three-dimensional morphology data h SEI(x, t), electrode stress field σ(x, t).

[0092] Electrochemical-mechanical coupling modeling includes lithium ion diffusion equation, SEI film growth morphology modeling, and electrode stress distribution model, specifically including:

[0093] (1) Lithium ion diffusion equation (Fick's second law), expressed as:

[0094]

[0095] Where c(x,t) is the lithium ion concentration field (mol / m 3 ), D eff (T) is the temperature-dependent effective diffusion coefficient (m 2 / s), R int (x,t) is the interface reaction consumption term (mol / (m 3 ·s)), originating from SEI growth and lithium insertion reaction;

[0096] (2) SEI film growth morphology modeling (based on Parabolic Kinetics), expressed as:

[0097]

[0098] Among them, h SEI (x, t) is the SEI film thickness (nm), k SEI is the SEI film growth rate constant (nm 2 / s), h0 is the initial SEI film thickness (nm);

[0099] (3) Electrode stress distribution model (thermal-electrical-mechanical coupling equation), expressed as:

[0100]

[0101] Where σ(x,t) is the equivalent stress at the electrode position x, E is the Young's modulus of the electrode, and α T is the thermal expansion coefficient (1 / K), β c is the concentration expansion coefficient (m 3 / mol).

[0102] The 3D state mapping module includes:

[0103] Normalization of state parameters: Normalize the lithium ion concentration field, SEI film thickness, electrode stress and temperature gradient to the interval [0,1], expressed as:

[0104]

[0105] Where F(x,t) is any original physical field variable, is the normalized physical variable, F min ,F max Respectively represent the minimum and maximum values ​​of the variable in the historical or sample range;

[0106] Calculation of health status indicators: Calculate health status indicators based on normalized variables;

[0107] Three-dimensional state matrix construction and output: The health status indicators are combined to form a three-dimensional state matrix, which is expressed as:

[0108]

[0109] Among them, H(x,t) is a three-dimensional health state matrix, which represents the health evolution state of the battery at different space and time points.

[0110] Health status indicators include:

[0111] Negative electrode expansion rate η meg (x,t), expressed as:

[0112]

[0113] in, is the normalized electrode strain, is the normalized lithium ion concentration, κ1, κ2 are empirical weighting coefficients, κ1∈[0.4,0.7], κ2∈[0.3,0.6];

[0114] Decomposition degree of electrolyte Expressed as:

[0115]

[0116] in, is the normalized temperature gradient, λ T is the temperature gradient response coefficient, controlling the slope of the curve, λ T ∈[2,5];

[0117] Interfacial reaction entropy change ΔS int (x,t), expressed as:

[0118]

[0119] in, is the normalized SEI film thickness, is the normalized initial film thickness (set to 1), γ is the interface structure entropy change sensitivity coefficient, γ∈[1,3].

[0120] The life degradation prediction module includes:

[0121] Health status matrix time series construction: Expand the three-dimensional health status matrix along the time dimension to construct the time series of each indicator;

[0122] Coupled degradation model calculation: Based on the multi-factor coupling characteristics of battery degradation, a weighted cumulative damage model is used to estimate the life consumption rate, which is expressed as:

[0123]

[0124] Where D(t) is the cumulative degradation as of time t (dimensionless), w1, w2, and w3 are the degradation weight factors of the health status indicators, satisfying w1+w2+w3=1, which are used to reflect the impact of each factor on the overall lifespan;

[0125] Remaining life and failure probability prediction: Based on the cumulative degradation amount D(t), calculate the battery remaining life RUL(t) and failure probability distribution P f (t), including:

[0126] (1) Remaining life estimation formula:

[0127] RUL(t)=T end -t,whereD(t)=D th ;

[0128] Among them, T end The first time the cumulative degradation reaches the threshold D th The corresponding time point, D th is the degradation critical threshold, which is set in the interval [0.85, 0.95];

[0129] (2) Failure probability distribution (using Weibull model):

[0130]

[0131] Among them, P f (t) is the cumulative failure probability at time t, λ is the scale parameter, and its value range is [800, 2000].

[0132] k is a shape parameter that reflects the failure rate trend and has a value range of [1.5, 3.5].

[0133] The model self-optimization module includes:

[0134] Failure risk perception: Receive the failure probability distribution P output by the life degradation prediction module f (t), within the set time window [t, t+Δt], calculate the local failure growth rate, which is expressed as:

[0135] ΔP f (t) = Pf (t+Δt)-P f (t);

[0136] Where ΔP f (t) is the local failure growth rate, P f (t) is the cumulative failure probability at time t, Δt is the time window step, in hours or cycles;

[0137] Dynamic parameter adjustment strategy generation: based on the failure growth rate ΔP f (t) Dynamically adjust the digital twin modeling parameters, and use the weighted adjustment coefficient ρ(t) to express the adjustment amplitude, which is expressed as:

[0138] ρ(t)=1+ξ·ΔP f (t);

[0139] Where ρ(t) is the parameter adjustment factor (dimensionless), which is used to amplify or reduce the model update frequency. ξ is the sensitivity coefficient, which controls the amplification ratio of the failure probability change to the parameter adjustment. The value range is ξ∈[5,10]. The key parameters of the model are dynamically updated according to ρ(t). The update rule is expressed as:

[0140] θ new =θ old ×ρ(t);

[0141] Among them, θ old ,θ new are the model parameters before and after the update respectively;

[0142] Optimize the sampling frequency command output: Synchronously adjust the sensor sampling frequency f of the data acquisition module based on the failure risk level s (t), the adjustment formula is expressed as:

[0143] f s (t) = f0 × (1 + λ s ΔP f (t));

[0144] Among them, f s (t) is the dynamically adjusted sampling frequency (Hz), f0 is the standard basic sampling frequency (Hz), λ s is the sampling frequency sensitivity coefficient, λ s ∈[2,5]; when the failure risk is high, the sampling frequency is increased to enhance the anomaly capture capability; when the failure risk is low, the sampling frequency is restored to the basic level to reduce energy consumption.

[0145] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Within the scope of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

Claims

1. Digital twin lithium battery real-time status monitoring system, characterized by: Includes the following modules: Multi-source heterogeneous data acquisition module: Configured with a distributed sensor array, it outputs multi-physics field raw data including dynamic values ​​of electrolyte viscosity, electrode strain waveforms, and local temperature gradients; Time-varying feature fusion module: receiving the multi-physics field raw data and generating a fusion feature vector with a unified timestamp through a spatiotemporal registration algorithm; Dynamic twin modeling module: Based on the fused feature vector, a cross-scale digital twin model including electrochemical-mechanical coupling effects is constructed to output real-time parameters of lithium ion concentration field, SEI film three-dimensional morphology, and electrode stress field; Three-dimensional state mapping module: converts the real-time parameters into battery health status indicators and outputs a three-dimensional state matrix including the negative electrode expansion rate, electrolyte decomposition degree, and interface reaction entropy change; Life degradation prediction module: Based on the three-dimensional state matrix, the battery remaining service life prediction value and failure probability distribution are calculated through a coupled degradation model; Model self-optimization module: dynamically adjusts the parameter update strategy of the digital twin model based on the failure probability distribution, and outputs the optimized sensor sampling frequency instruction to the data acquisition module.

2. The digital twin lithium battery real-time status monitoring system according to claim 1 is characterized in that: The multi-source heterogeneous data acquisition module includes: Distributed sensor deployment: sensor arrays are deployed in key locations of the lithium battery structure, including micro-electroviscosity sensor arrays in the electrolyte area, strain gauge sensors on the electrode surface, and MEMS temperature sensors in the gaps between battery cell layers. Multi-physical quantity signal acquisition: Various sensors synchronously collect dynamic values ​​of electrolyte viscosity, electrode strain waveform, and local temperature gradient; Time-scale unification of multi-source data: synchronize the timestamps and match the spatial labels of the collected data to construct a multi-physics field raw data set including viscosity, strain and temperature gradient information as input for the subsequent feature fusion module.

3. The digital twin lithium battery real-time status monitoring system according to claim 2 is characterized in that: The unified time scale of multi-source data includes: Timestamp synchronization processing: synchronize the sampling time of electrolyte viscosity value, electrode strain waveform, and local temperature gradient, and use dynamic time warping algorithm on the original sampling sequence to align them on the common time axis; Spatial label matching: Map the sampling points to a unified spatial reference system based on the physical location of the sensor and set a unified spatial label set; Construction of multi-physics field original data set: Construct a fusion data matrix of three types of physical quantities in a unified space-time coordinate system.

4. The digital twin lithium battery real-time status monitoring system according to claim 3 is characterized in that: The time-varying feature fusion module includes: Unified timestamp inheritance processing: Unified timestamp inheritance processing, time alignment of multi-source physical quantities, ensuring that subsequent processing is performed under the same time base to avoid repeated calculations; Unified spatial registration and interpolation: The aligned physical quantity data are mapped to a unified spatial coordinate set, and the Gaussian kernel interpolation algorithm is used for spatial registration; Fusion feature vector generation: The physical quantities after spatiotemporal registration are constructed into a fusion feature vector.

5. The digital twin lithium battery real-time status monitoring system according to claim 4 is characterized in that: The dynamic twin modeling module includes: Input feature-driven modeling initialization: Receive the fused feature vector output by the time-varying feature fusion module as the model input, drive the joint modeling process of electrochemical and mechanical behaviors, and establish the foundation for the coupling of the three physical fields of electricity, heat, and force; Electrochemical-mechanical coupling modeling: Based on a set of coupled partial differential equations, it simulates the key internal physical processes of lithium batteries during operation. Real-time output of state parameters: Based on the numerical solution of the partial differential coupled equations, three types of lithium battery state parameters are output at a set time step, including lithium ion concentration field, SEI film 3D morphology data, and electrode stress field.

6. The digital twin lithium battery real-time status monitoring system according to claim 5, characterized in that: The electrochemical-mechanical coupling modeling includes lithium ion diffusion equation, SEI film growth morphology modeling, and electrode stress distribution model.

7. The digital twin lithium battery real-time status monitoring system according to claim 6, characterized in that: The three-dimensional state mapping module includes: State parameter normalization processing: normalizing the lithium ion concentration field, SEI film thickness, electrode stress and temperature gradient to the interval [0,1]; Calculation of health status indicators: Calculate health status indicators based on normalized variables; Three-dimensional state matrix construction and output: health status indicators are combined to form a three-dimensional state matrix.

8. The digital twin lithium battery real-time status monitoring system according to claim 7, characterized in that: The health status indicators include the negative electrode expansion rate, the electrolyte decomposition degree and the interface reaction entropy change.

9. The digital twin lithium battery real-time status monitoring system according to claim 8, characterized in that: The life degradation prediction module includes: Health status matrix time series construction: Expand the three-dimensional health status matrix along the time dimension to construct the time series of each indicator; Coupled degradation model calculation: Based on the multi-factor coupling characteristics of battery degradation, a weighted cumulative damage model is used to estimate the life consumption rate; Remaining life and failure probability prediction: Calculate the remaining battery life and failure probability distribution based on the accumulated degradation.

10. The digital twin lithium battery real-time status monitoring system according to claim 9, characterized in that: The model self-optimization module includes: Failure risk perception: Receives the failure probability distribution output by the life degradation prediction module and calculates the local failure growth rate within a set time window; Dynamic parameter adjustment strategy generation: Dynamically adjust digital twin modeling parameters based on failure growth rate, using weighted adjustment coefficients to express the adjustment range; Optimize sampling frequency instruction output: Synchronously adjust the sensor sampling frequency of the data acquisition module based on the failure risk level.

Images