Fault diagnosis method based on hybrid capacitor electrochemical model

Abstract

The embodiment of the invention discloses a fault diagnosis method based on a hybrid capacitor electrochemical model, and relates to the technical field of hybrid capacitor health management, and the method comprises the steps: obtaining the experimental data of a hybrid capacitor in a plurality of degradation stages; establishing an electrochemical model, and selecting a plurality of mechanism parameters as a to-be-identified set; performing parameter identification on the to-be-identified set by taking an experimental voltage curve as a reference and adopting an optimization strategy taking a terminal voltage error as a target function to obtain a mechanism parameter sequence evolved along with a degradation stage; extracting offset characteristics of the mechanism parameter sequence; establishing a mapping relationship between the offset feature and a preset typical fault mode to identify a fault type; a weighted degradation index is constructed based on the offset features, and the fault severity is quantified; the method can realize interpretable diagnosis of the degradation mechanism of the hybrid capacitor, has the advantages of high precision, strong generalization, irrelevance with working conditions and the like, and can be used for health management and safety early warning of the hybrid capacitor and an energy storage system.

Description

Technical Field

[0001] This invention relates to the field of power system technology, and in particular to a fault diagnosis method based on a hybrid capacitor electrochemical model. Background Technology

[0002] Frequency stability of a power system is fundamental to ensuring the safe operation of the power grid. System inertia, as a core physical quantity that resists frequency changes, is crucial for maintaining frequency stability. In traditional power systems, rotational inertia is mainly provided by the rotating components of synchronous generator sets. When power disturbances occur in the system, these components can spontaneously release or absorb kinetic energy, thereby slowing down the rate of frequency change.

[0003] However, with the large-scale integration of new energy power generation forms such as wind power and photovoltaics into the grid via power electronic converters, the composition of the power system has undergone fundamental changes. These new energy units themselves do not possess the inertial response capability of traditional synchronous units, resulting in a significant decrease in the equivalent inertia level of the entire power system. Consequently, the frequency changes of the system are more drastic when subjected to disturbances, increasing the operational risks of the power grid.

[0004] To address this challenge, it is urgent to develop system inertia assessment methods that take into account new energy sources. Current inertia assessment methods mostly consider conventional power sources such as hydropower and thermal power, while giving less consideration to the inertia constants of new energy power plants such as wind power and photovoltaic power, and the consideration is not accurate enough. This makes it impossible to accurately assess the inertia constants of high-proportion new energy systems, which may lead to potential frequency instability risks in the system. Summary of the Invention

[0005] The main objective of this invention is to provide a fault diagnosis method based on a hybrid capacitor electrochemical model to solve the problems of current methods being unable to reveal internal degradation mechanisms, being sensitive to operating conditions, and relying on a large number of fault samples, thereby achieving early, interpretable, and condition-independent mechanism-level fault identification and severity quantification.

[0006] To achieve the above objectives, the first aspect of this application provides a fault diagnosis method based on a hybrid capacitor electrochemical model, the method comprising: Experimental data were obtained for hybrid capacitors at multiple degradation stages, including capacitance, internal resistance, electrochemical impedance spectroscopy, and voltage-time curves. An electrochemical model incorporating solid-phase diffusion, interfacial reactions, and electrolyte transport was established, and several degradation-sensitive mechanistic parameters were selected as the set to be identified. Based on the experimental voltage curve, an optimization strategy with terminal voltage error as the objective function is adopted to identify the parameters of the set to be identified, and obtain the sequence of mechanism parameters that evolve with the degradation stage. Extract the offset features of the mechanism parameter sequence; The offset features are mapped to preset typical fault modes to identify fault types; A weighted degradation index is constructed based on the aforementioned offset features to quantify the severity of the fault.

[0007] A second aspect of this application provides a fault diagnosis device based on a hybrid capacitor electrochemical model, comprising: The data acquisition module is used to acquire experimental data of hybrid capacitors at multiple degradation stages, including capacitance, internal resistance, electrochemical impedance spectroscopy, and voltage-time curves. The model building module is used to establish an electrochemical model that includes solid-phase diffusion, interfacial reactions and electrolyte transport, and selects multiple degradation-sensitive mechanism parameters as a set to be identified. The parameter identification module is used to identify the parameters of the set to be identified by using the experimental voltage curve as a reference and an optimization strategy with the terminal voltage error as the objective function, so as to obtain the mechanism parameter sequence that evolves with the degradation stage. The feature extraction module is used to extract the offset features of the mechanism parameter sequence; The fault mapping module is used to establish a mapping relationship between the offset features and preset typical fault modes in order to identify the fault type; The fault quantification module is used to construct a weighted degradation index based on the offset features to quantify the severity of the fault.

[0008] A third aspect of this application provides an electronic device including a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform steps as described in the first aspect and any possible implementation thereof.

[0009] A fourth aspect of this application provides a computer-readable storage medium, wherein when a computer program is executed by a processor, the processor performs the steps of the first aspect and any possible implementation thereof.

[0010] This application provides a fault diagnosis method based on an electrochemical model of a hybrid capacitor. It acquires experimental data of the hybrid capacitor at multiple degradation stages, including capacitance, internal resistance, electrochemical impedance spectroscopy, and voltage-time curves. An electrochemical model incorporating solid-phase diffusion, interfacial reactions, and electrolyte transport is established, and multiple degradation-sensitive mechanistic parameters are selected as a set to be identified. Using the experimental voltage curves as a benchmark, an optimization strategy with terminal voltage error as the objective function is employed to identify the parameters in the set to be identified, obtaining a sequence of mechanistic parameters evolving with each degradation stage. The shift features of the mechanistic parameter sequence are extracted. A mapping relationship is established between the shift features and preset typical fault modes to identify the fault type. A weighted degradation index is constructed based on the shift features to quantify the fault severity. This method enables interpretable diagnosis of the degradation mechanism of hybrid capacitors, offering advantages such as high accuracy, strong generalization, and independence from operating conditions. It can be widely applied to the health management of power batteries, energy storage systems, drones, and power systems in extreme environments. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application 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 only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] in: Figure 1 A schematic flowchart illustrating a fault diagnosis method based on a hybrid capacitor electrochemical model provided in this application embodiment; Figure 2 This is a schematic diagram of the degradation mechanism of a hybrid capacitor provided in an embodiment of this application; Figure 3 A schematic diagram of the structure of a fault diagnosis device based on a hybrid capacitor electrochemical model provided in this application embodiment; Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0013] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0014] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0015] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0016] The embodiments of this application are described below with reference to the accompanying drawings.

[0017] Figure 1 This is a schematic flowchart illustrating a fault diagnosis method based on a hybrid capacitor electrochemical model, provided as an embodiment of this application. Figure 1 As shown, the method includes: 101. Obtain experimental data of hybrid capacitors at multiple degradation stages, including capacitance, internal resistance, electrochemical impedance spectroscopy, and voltage-time curves.

[0018] Specifically, the target hybrid capacitor can be subjected to capacitance testing, internal resistance measurement, voltage-time charge-discharge sampling, electrochemical impedance spectroscopy testing, intermittent titration (GITT) testing, or other experimental tests that can be used to characterize degradation characteristics at multiple cycle nodes to obtain data such as terminal voltage, operating current, temperature, and impedance to characterize the state of the hybrid capacitor.

[0019] In one alternative implementation, the acquisition of experimental data on the hybrid capacitor at multiple degradation stages includes: The performance parameters of the above hybrid capacitor were tested at multiple turns. These performance parameters include capacitance, coulombic efficiency, internal resistance, charge transfer resistance, diffusion coefficient, heat generation rate, and thermal conductivity / specific heat. By correlating the above performance parameters with the corresponding fault indices for each cycle number, a degradation fault dataset for the hybrid capacitor is constructed.

[0020] For example, performance parameters of a hybrid capacitor can be tested at the 1st, 2nd, 5th, 10th, 20th, 50th, 100th, 200th, 500th, 1000th, 2000th, 3000th, and so on. These parameters include capacitance, coulombic efficiency, internal resistance, charge transfer resistance (Rct), Warburg coefficient, diffusion coefficient, heat generation rate, and thermal conductivity / specific heat. By correlating these failure indicators with the number of cycles, a degradation failure dataset for the hybrid capacitor can be constructed.

[0021] 102. Establish an electrochemical model that includes solid-phase diffusion, interfacial reaction and electrolyte transport, and select multiple mechanistic parameters that are sensitive to degradation as the set to be identified.

[0022] Specifically, based on the electrochemical reaction mechanism of hybrid capacitors, an electrochemical model incorporating processes such as solid-phase diffusion, electrolyte transport, and interfacial charge transfer reactions can be established. Multiple (e.g., at least 6–10) degradation-sensitive model parameters can be selected as parameters to be identified, forming a set of parameters to be identified. These may include solid-phase diffusion coefficient, maximum solid concentration, porosity, electrolyte diffusion coefficient, charge transfer rate constant, exchange current density, and side reaction rate constant. Further optionally, the electrochemical model may also include the SEI growth process; in this case, the parameters to be identified may also include SEI thickness or its equivalent impedance parameters.

[0023] In an optional implementation, step 102 proposes a method of mapping the electrochemical model parameters to be identified (including: solid diffusion coefficients Dsn, Dsp; maximum solid concentrations csmax_n, csmax_p; active material volume fraction εs; electrode porosity ε; exchange current density i0; charge transfer rate constant k; electrolyte conductivity κ; electrolyte diffusion coefficient De; transport number t+; SEI film thickness δ_SEI; SEI conductivity_SEI; side reaction rate constant k_SEI) to the labels of hybrid capacitor degradation and failure, as an intermediate basis for hybrid electrical appliance fault diagnosis.

[0024] 103. Based on the experimental voltage curve, an optimization strategy with terminal voltage error as the objective function is adopted to identify the parameters of the above set to be identified, and the sequence of mechanism parameters that evolve with the degradation stage is obtained.

[0025] Specifically, using the experimental voltage curve as a reference and the model simulation voltage curve as the output, an objective function including the root mean square error (RMSE) is constructed. The model parameters can be identified by combining global search and local optimization to obtain the set of electrochemical model parameters corresponding to different cycle stages.

[0026] 104. Extract the offset features of the above mechanism parameter sequence.

[0027] Specifically, time series analysis can be used to calculate the parameter change rate, standardized residual, offset magnitude, and differences from the normal degradation model, thereby forming the offset feature vector of electrochemical mechanism parameters.

[0028] In step 104, a method can be used to extract comprehensive offset features, including parameter standardized residuals, parameter change slopes, change rates within short-term windows, and cross-parameter coupling features, to process fault data. The aforementioned cross-parameter coupling features are used to characterize multi-mechanism collaborative degradation.

[0029] In this embodiment, the objective function can be based on the terminal voltage RMSE, and multi-condition weighting and regularization terms can be introduced to measure the deviation between the model output and the experimental data, making the identification results more robust.

[0030] 105. Establish a mapping relationship between the above-mentioned offset features and preset typical fault modes to identify fault types.

[0031] Among them, the typical failure modes mentioned above include abnormal SEI growth, loss of active materials, electrolyte degradation, and combinations thereof.

[0032] After obtaining the offset features, a one-to-one mapping relationship is established between model parameters and typical failure modes. Specifically, the offset features of the obtained model parameters such as solid phase, interface, electrolyte, and structural geometry can be mapped to typical failure mechanisms such as SEI growth, active material loss, electrolyte degradation, and multi-mechanism synergistic degradation. Subsequent fault diagnosis and fault type identification can then be performed: based on the parameter offset patterns, fault determination rules are used to identify whether hybrid capacitors exhibit abnormal SEI growth, active material loss, electrolyte degradation, or multi-mechanism synergistic degradation, and the corresponding fault type is output.

[0033] The above-mentioned fault types may include: SEI abnormal growth; Loss of positive or negative electrode active materials; Electrolyte deterioration; diffusion-restricted degradation; Pore ​​blockage-type degradation; Multiple mechanisms of synergistic degradation.

[0034] 106. Based on the above offset characteristics, a weighted degradation index is constructed to quantify the severity of the fault.

[0035] Specifically, based on the parameter offset magnitude, residuals, and trends, a fault degree quantification index is constructed to quantify the degree of degradation of hybrid capacitors. For example, it can be classified into mild, moderate, and severe levels, or other levels or scores, without any restrictions here.

[0036] Optionally, in fault diagnosis and identification, the contribution of multiple degradation factors can be quantitatively assessed by learning the weight of each feature to the overall degradation degree, and a degradation degree predictor can be constructed using an Elastic Net regression model.

[0037] The fault diagnosis method based on the electrochemical model of a hybrid capacitor in this application embodiment acquires experimental data of the hybrid capacitor at multiple degradation stages, including capacitance, internal resistance, electrochemical impedance spectroscopy, and voltage-time curves. An electrochemical model incorporating solid-phase diffusion, interfacial reactions, and electrolyte transport is established, and multiple degradation-sensitive mechanistic parameters are selected as a set to be identified. Using the experimental voltage curve as a benchmark, an optimization strategy with terminal voltage error as the objective function is employed to identify the parameters of the set to be identified, obtaining a sequence of mechanistic parameters evolving with each degradation stage. The offset features of the mechanistic parameter sequence are extracted. A mapping relationship is established between the offset features and preset typical fault modes to identify the fault type. A weighted degradation index is constructed based on the offset features to quantify the fault severity. This method enables interpretable diagnosis of the degradation mechanism of hybrid capacitors, offering advantages such as high accuracy, strong generalization, and independence from operating conditions. It can be widely applied to the health management of power batteries, energy storage systems, drones, and power systems in extreme environments.

[0038] To more clearly illustrate the methods and effects of the embodiments of this application, the following description is based on experimental data.

[0039] 1. A research method based on experiments to obtain degradation parameters of hybrid capacitors, providing real and detailed data.

[0040] To establish the mapping relationship between the degradation mechanism of hybrid capacitors and electrochemical model parameters, and to further support fault diagnosis methods based on model parameters, this application systematically measured and analyzed the key degradation parameters of hybrid capacitors throughout their entire life cycle. Solid-phase, electrolyte, interface, and system-level performance parameters were obtained through various experimental methods, constructing a complete degradation parameter database to provide fundamental data support for subsequent mechanism analysis and diagnostic algorithms.

[0041] 1.1 Degradation Period Sampling Strategy (1) To capture the continuous evolution of the degradation process of hybrid capacitors, this application measures parameters at typical cycle nodes starting from the activation stage of the hybrid capacitor, including the 1st, 2nd, 5th, 10th, 20th, 50th, 100th, 200th, 500th, 1000th, 2000th, 3000th, etc. lifetime stages. By constructing a time series of degradation parameters through multi-timescale sampling, the long-term evolution and abrupt change characteristics of various mechanistic parameters can be completely recorded.

[0042] (2) Based on the parameter degradation law obtained from the test, the solid diffusion coefficient D_s, porosity ε, maximum solid concentration c_smax, and transfer number t are artificially changed. + To simulate the degradation process of a hybrid capacitor, data is generated through parameters. Improvements are needed based on this, requiring: in addition to generating normal degradation data, data for different fault types should be added; parameter sample data is needed to generate time-voltage data, providing data for parameter identification. That is, both time-voltage and time-voltage data sample data should be generated; the parameter sample data structure should be set as described below.

[0043] (3) The sample structure consists of three parts: meta-information, features and labels.

[0044] I. Meta-information includes: cell_id: cell number (can be HSC-001, HSC-0002, etc.); cycle_index: cycle number (e.g., 0, 2, 5, 10, 20, 50, 100, 200, 500, 1000, 2000, 3000, 4000, etc.); T_amb: ambient temperature (e.g., 15℃, 25℃, etc.); rate: rate of operation (e.g., 1C, 2C, 5C, etc.).

[0045] II. Features include: (1) Obtainable electrochemical parameters, such as: positive electrode diffusion coefficient (Dsp); negative electrode diffusion coefficient (Dsn); positive electrode porosity (eps_p); negative electrode porosity (eps_n); maximum solid lithium concentration of positive electrode (csmax_p); maximum solid lithium concentration of negative electrode (csmax_n); positive electrode conductivity (kp); negative electrode conductivity (kn); transfer number (t_plus). (2) In addition, some relative change and residual features: z_Dsp, z_Dsn, z_eps_p, z_eps_n, z_csmax_p, z_csmax_n, z_kp, z_kn, z_tplus, that is: the standardized residuals relative to the normal degradation model. For example, the rate of change (slope) in the last 100 cycles reflects whether the degradation rate suddenly accelerates.

[0046] III. The fault type label is defined as a single-label multi-classification problem. For example, the `fault_type` field can take values ​​such as: Normal (normal, no obvious fault); LAM_p (dominated by loss of positive electrode active material); LAM_n (dominated by loss of negative electrode active material); Diffusion_Limited (diffusion-limited); Pore_Clogging (pore / structure blockage); Electrolyte_Degradation (abnormal electrolyte / ion migration). A severity label is also added: `fault_severity`, using a continuous value of 0–1 to indicate the severity, further labeled as normal, mild, moderate, and severe.

[0047] 1.2 Important Degradation Parameters of Electrochemical Models Solid-state parameters primarily reflect the degradation of active materials, reversible lithium loss, diffusion restriction, and changes in particle structure, and are an important component of electrochemical models. Electrolyte and interface parameters determine ion migration, charge transfer, and SEI growth behavior, and their changes directly reflect the degree of kinetic degradation.

[0048] (1) Maximum solid concentration (positive / negative electrode). The maximum solid concentration is inferred from the dead lithium content / remaining capacity, which reflects the degree of dissipation of the available lithium inventory in the electrode with cycling.

[0049] (2) Initial solid phase concentration (positive / negative electrode). Calculated based on the dead lithium content and parameter identification results, reflecting the formation of SEI and the initial capacity loss.

[0050] (3) Porosity (ε). The electrode microstructure was characterized using scanning electron microscopy (SEM), or the change in porosity with cycling was inverted using parameter identification methods. The decrease in porosity is usually closely related to microstructure compaction, pore blockage, and deposition of lithium by-products.

[0051] (4) Active material particle size (positive / negative electrode). The change in electrode particle size is measured by SEM to reflect the structural deterioration process of active material, such as crushing, breaking or expansion.

[0052] (5) Diffusion coefficient (Ds, positive / negative electrode). The solid-phase diffusion coefficient is obtained by GITT (intermittent titration) experiment and is an important parameter describing the migration ability of lithium ions in solid particles. Its decrease usually corresponds to diffusion restriction or increased polarization caused by SEI thickening.

[0053] (6) Electrolyte conductivity (κ). It is obtained by fitting the resistance in the low-frequency range of electrochemical impedance spectroscopy (EIS) and reflects electrolyte degradation, salt concentration changes and decreased migration ability.

[0054] (7) Specific surface area (as). Extracted by parameter identification method, it reflects the change of interfacial active area and is closely related to SEI growth, dendrite formation and interfacial reaction rate.

[0055] (8) Charge transfer impedance (Rct). Obtained by mid-frequency arc fitting in EIS, it is a key characterization parameter for interfacial reaction kinetics. An increase in Rct usually indicates SEI thickening, electrode interface deactivation, or accelerated electrolyte decomposition.

[0056] 1.3 System-level degradation characteristic testing System-level parameters can provide information on the macroscopic performance degradation of hybrid capacitors, helping to bridge the gap between mechanistic parameters and performance indicators.

[0057] (1) Internal resistance test. The internal resistance of the positive and negative terminals is measured directly by an internal resistance tester.

[0058] (2) HPPC (Hybrid Pulse Power Characteristic) test. The change in DC internal resistance (DCIR) and polarization curve are obtained through multi-pulse testing. The increase in DCIR corresponds to dynamic degradation and interface aging, which is the standard method for monitoring dynamic degradation.

[0059] (3) OCV–SOC curve measurement. The open-circuit voltage curves of different degradation nodes are measured to determine whether the voltage plateau has shifted. This shift may originate from changes in the true equilibrium potential.

[0060] (4) Temperature and pressure monitoring. Temperature and pressure changes are recorded synchronously during the cycle, which can be used to detect side reactions, gas generation and thermal stability degradation, and is an important data source for safety-related diagnostics.

[0061] (5) Electrochemical Impedance Spectroscopy and Parameter Fitting. EIS measurements were performed under different cycle counts. By fitting an equivalent circuit model, the following parameters could be extracted: intrinsic resistance (Rs), interfacial charge transfer impedance (Rct), double-layer capacitance / constant phase angle element parameters, Warburg diffusion impedance, and interfacial SEI characteristic impedance. These parameters are helpful in analyzing the core mechanisms of SEI growth, electrolyte degradation, polarization intensification, and interfacial deactivation.

[0062] (6) Application of GITT test. The GITT test can simultaneously obtain: solid diffusion coefficient Ds; true equilibrium open circuit potential Ueq.

[0063] Compared to ordinary OCV testing, GITT can effectively distinguish between voltage shifts caused by polarization and changes in the true equilibrium potential, making it an important experimental method for constructing accurate model parameters.

[0064] Through the aforementioned experimental methods, this application constructed a "full-parameter degradation database" covering solid-phase structure, electrolyte performance, interfacial dynamics, and system-level performance. These parameters not only highly correspond to the mechanistic parameters in electrochemical models (such as the P2D model), but also clearly reflect the evolution patterns of different degradation mechanisms, including active material loss, restricted lithium diffusion, decreased electrolyte conductivity, SEI thickening, and enhanced polarization. These experimental results lay a solid data foundation for subsequent fault type identification, fault severity quantification, and lifetime prediction based on electrochemical model parameters.

[0065] 2. Establish a one-to-one mapping relationship between electrochemical mechanism model parameters and typical failure modes. This step in this application aims to establish a systematic correspondence between externally measurable degradation parameters, internal electrochemical mechanism model parameters, and typical failure modes of hybrid capacitors, providing a theoretical basis for subsequent fault diagnosis algorithms, degradation model correction strategies, and health status estimation. First, the degradation phenomena and mechanisms of hybrid capacitors are analyzed. Second, the problems reflected by experimentally measurable degradation parameters are explained. Then, the correspondence between degradation parameters and electrochemical model parameters is established, ultimately forming a mapping of the impact of hybrid capacitor degradation on mechanism model parameters.

[0066] 2.1 Degradation Phenomenon and Mechanism of Hybrid Capacitors Hybrid capacitors are subjected to a combination of physical and chemical processes during cycling, which causes their performance to gradually degrade. Figure 2 The multidimensional degradation path of hybrid capacitors from the structural layer, interface layer and electrolyte layer is shown. Table 1 shows the degradation mechanism of aged hybrid capacitors and summarizes the core degradation mechanism and its corresponding performance.

[0067]

[0068] Table 1 From a mechanistic perspective, the degradation of hybrid capacitors is mainly caused by four types of mechanisms: (1) SEI growth: As cycling proceeds, the SEI film continuously thickens, consuming active lithium and reducing the amount of recyclable lithium, thus causing capacity decay. At the same time, the thickening of the film obstructs the interfacial migration channels between lithium ions and electrons, leading to an increase in interfacial impedance.

[0069] (2) Loss of active material: Particle breakage, pulverization or shedding leads to a reduction in effective active material, enhanced electrode polarization, manifested as capacity decay and decreased rate capability; damage to the material structure can also cause interruption of electron paths, resulting in increased internal resistance.

[0070] (3) Electrolyte deterioration: Electrolyte decomposition or increased viscosity will reduce ionic conductivity and increase ohmic impedance, especially limiting the energy output capability of hybrid capacitors at high rates.

[0071] (4) Synergistic effect of multiple mechanisms: As the processes of SEI growth, material degradation and electrolyte aging are superimposed, the capacity and power performance of hybrid capacitors will decrease simultaneously, eventually reaching the end-of-life (EOL) state where the capacity is less than 80% or the internal resistance is more than twice the initial value.

[0072] The above mechanisms constitute the physical basis for the performance changes of hybrid capacitors, providing a basis for establishing parameter-fault mode mapping from the perspective of mechanism modeling.

[0073] 2.2 Problems revealed by experimentally measurable degradation parameters In actual testing and operational monitoring, a series of degradation parameters can be obtained through methods such as capacity testing, DC internal resistance measurement, electrochemical impedance spectroscopy (EIS) testing, and thermal response testing. Table 2 is a comparison table of degradation parameters for hybrid capacitors, which systematically summarizes these parameters and their physical meanings.

[0074]

[0075] Table 2 (1) Electrical performance parameters reflect the overall health status. Capacitance reflects the energy storage capacity of hybrid capacitors, and its decline with cycling directly reflects the degradation of active materials or reversible lithium, making it the most common SOH indicator. Coulombic efficiency below 100% indicates that side reactions are continuing and can be used to evaluate the SEI growth rate or the degree of electrolyte decomposition. Decreased energy efficiency indicates increased power loss due to increased internal resistance.

[0076] (2) Impedance and kinetic parameters reveal interface degradation. An increase in DC internal resistance (DCIR) indicates that ohmic impedance and polarization impedance increase simultaneously, which is a key indicator for rapid screening and online monitoring of SOH on the production line. An increase in charge transfer resistance (Rct) reflects a deterioration in interfacial reaction kinetics, which is related to the thickening of SEI or the reduction of active material. Changes in Warburg impedance reflect changes in lithium-ion diffusion capacity and can indicate internal structural defects or pore blockage in particles.

[0077] (3) The degradation of structural property parameters is caused by the increase in SEI thickness, which corresponds to the increase in interfacial impedance; the decrease in the utilization rate of active materials indicates that the particles are broken or fall off; the formation of lithium dendrites may lead to irreversible capacity loss or even short circuit.

[0078] (4) Thermal parameters reflect the decline in thermal stability. An increase in the heat generation rate indicates increased I²R loss or enhanced exothermic side reactions; changes in thermal conductivity / specific heat reflect the heat transfer path or the decline in the thermal properties of the material.

[0079] These measurable parameters provide direct evidence for determining the source of degradation and assessing health levels, and can also serve as inputs for parameter identification in mechanistic models.

[0080] 2.3 Correspondence between degradation characteristic parameters and electrochemical mechanism model parameters To achieve an interpretable mapping from measurable degradation data to internal mechanisms, it is necessary to establish a correspondence between degradation parameters and electrochemical model parameters. Table 3 shows one such mapping table between degradation parameters and mechanistic model parameters.

[0081]

[0082] Table 3 As shown in Table 3, different types of degradation characteristics can be mapped to changes in parameters such as solid-phase diffusion, interfacial reaction, and electrolyte transport in the electrochemical mechanism model.

[0083] (1) Capacity decay corresponds to changes in the relevant parameters of the active material in the electrochemical model. The volume fraction of the active material εs decreases; the maximum solid concentration cs_max decreases; and the electrode structural parameters (such as thickness L) change. This reflects a reduction in the effective material that can participate in the lithium intercalation reaction.

[0084] (2) A low coulombic efficiency corresponds to an increase in the side reaction rate parameter k_SEI. CE < 100% indicates that the side reaction continues, which will lead to: reversible lithium loss; an increase in the side reaction rate constant (k_SEI), corresponding to continuous growth of the SEI film.

[0085] (3) An increase in DCIR corresponds to changes in ohmic resistance and SEI-related parameters. Ohmic resistance RΩ increases; SEI film resistance R_film increases. This indicates that the electrolyte conductivity decreases or the SEI film becomes thicker.

[0086] (4) The increase in the mid-frequency semicircle of EIS corresponds to the degradation of the charge transfer process. The charge transfer resistance Rct increases; the exchange current density i0 decreases; and the interface reaction rate constant k decreases. This reflects a slowdown in interface kinetics.

[0087] (5) The change in Warburg slope corresponds to a decrease in solid / liquid diffusion parameters. The solid diffusion coefficient Ds decreases; the electrolyte diffusion coefficient De decreases. This indicates that diffusion is more restricted.

[0088] (6) Increased heat generation corresponds to the combined effects of RΩ, Rct, and k_SEI. As the internal impedance increases and the side reactions intensify, the heat generation of the hybrid capacitor also increases.

[0089] The established correspondence realizes a one-to-one mapping from external observation indicators to internal parameters of the mechanism model, providing a quantitative basis for electrochemical fault diagnosis.

[0090] 2.4 Systemic effects of hybrid capacitor degradation on electrochemical mechanism model parameters Based on the above mapping relationship, the systematic influence of hybrid capacitor degradation on various parameters in the electrochemical model can be summarized as shown in Table 4.

[0091]

[0092] Table 4 Table 4 illustrates the systematic impact of degradation on the parameters of the mechanistic model. The effects can be summarized from aspects such as the solid phase, electrolyte, interface, thermal properties, and geometry.

[0093] (1) Solid-phase related parameters. The decrease in maximum solid concentration cs_max indicates that the loss of active material has led to a decrease in lithium storage capacity; the decrease in solid diffusion coefficient Ds indicates that particle cracks and pore blockage have restricted lithium intercalation diffusion; the decrease in solid conductivity σs indicates that material pulverization affects the electron transport path.

[0094] (2) Liquid-related parameters. A decrease in the electrolyte diffusion coefficient De indicates that the high viscosity of the electrolyte restricts ion diffusion; a decrease in the ionic conductivity κ indicates that electrolyte decomposition or deposit accumulation blocks ion migration; a decrease in the transport number t+ reflects the reduction of Li+ content in the electrolyte.

[0095] (3) Interface dynamic parameters. A decrease in exchange current density i0 and charge transfer rate constant k indicates that the interface reaction is slower; an increase in SEI thickness δ_SEI and a decrease in SEI conductivity κ_SEI represent a significant increase in interface impedance; an increase in side reaction rate constant k_SEI represents an intensification of side reactions.

[0096] (4) Thermal parameters. Reaction entropy coefficient U / Changes in temperature (T) affect the OCV–T curve; a decrease in heat capacity (Cp) and thermal conductivity (k) reflects electrode structure degradation and a decline in thermal management capability.

[0097] (5) Geometric parameters. Decreased porosity ε leads to blockage of reaction deposits, resulting in a reduction of ion transport channels; decreased specific surface area A leads to breakage or detachment of reaction particles, resulting in a decrease in the effective area of ​​the interface; changes in electrode thickness L lead to expansion, changes in compaction, or structural collapse due to long-term reaction cycles.

[0098] The changes in these parameters systematically reveal the degradation trajectory of hybrid capacitors in terms of capacitance decay, impedance increase, diffusion limitation, interface degradation, and decreased thermal stability, achieving a comprehensive characterization at the mechanistic level. A complete logical chain from degradation observation to mechanistic parameters to failure modes is established. This provides a complete theoretical foundation and data support for the typical failure mode identification method based on model parameter degradation trajectory proposed in this application.

[0099] 3. Methods for constructing and identifying parameters of electrochemical mechanism models To obtain the electrochemical mechanism model parameters corresponding to the degradation state of hybrid capacitors, this application further constructs a method for identifying electrochemical model parameters based on traditional optimization and least squares principles. Given an electrochemical mechanism model (P2D model), this method compares multi-condition experimental data with model simulation results to inversely deduce key parameters such as solid-phase diffusion, interfacial reactions, electrolyte transport, and structural geometry, thereby obtaining a set of parameters that reflects the degree of degradation of the hybrid capacitor.

[0100] 3.1 Parameter Selection and Identification Dimensions In electrochemical mechanism models, a large number of parameters can be identified. Considering the model's identifiability, computational complexity, and relevance to degradation mechanisms, this application selects 6–10 core mechanism parameters sensitive to degradation as identification targets. These include, but are not limited to: (1) Solid-phase related parameters: negative electrode solid-phase diffusion coefficient Dsn (can be in the form of log10Dsn), positive electrode solid-phase diffusion coefficient Dsp, negative electrode maximum solid-phase concentration csn_max, positive electrode maximum solid-phase concentration csp_max.

[0101] (2) Interface kinetic parameters: negative electrode exchange current density i0,n or reaction rate constant kn; positive electrode exchange current density i0,p or reaction rate constant kp.

[0102] (3) Electrolyte and structural parameters: effective conductivity κ of electrolyte eff Electrolyte diffusion coefficient De; Electrode porosity εe or active material volume fraction εs.

[0103] In practical implementation, the above parameters can be combined and selected according to the model form and experimental methods, for example, to identify a 6-dimensional parameter set: Θ={log 10 Dsn, log 10 The dimensionality can be expanded to 8-10 dimensions, such as Dsp, csn_max, csp_max, i0,n,i0,p}, to improve the ability to characterize degradation patterns.

[0104] 3.2 Experimental Design and Data Acquisition To ensure parameter identifiability, this application preferably employs a multi-condition joint identification strategy, including but not limited to: (1) Multi-rate constant current / constant current-constant voltage charge and discharge test: low rate (such as 0.5C, 1C) is used to highlight the characteristics of equilibrium state and diffusion process; medium and high rate (such as 2C, 5C) is used to highlight polarization and kinetic constraints.

[0105] (2) Excitation of multiple SOC ranges: Select several representative ranges (such as 10–20%, 40–60%, 80–90%) within the 0–100% SOC range to ensure that the responses on both sides of the electrode at different lithiation levels are fully excited.

[0106] (3) Optional electrochemical impedance spectroscopy (EIS) test: used to assist in constraining ohmic impedance, Rct and Warburg related parameters, and improve the reliability of identification results.

[0107] The main data obtained from the above tests include the time-series terminal voltage V. exp The current excitation I(tk), temperature T(tk), etc., are used as inputs and calibration objects for subsequent identification.

[0108] 3.3 Electrochemical Mechanism Model and Simulation Output In the parameter identification process, an electrochemical mechanism model with the parameter Θ to be identified is used: M:{I(t),T(t),Θ} Vsim(tk;Θ) That is, under given current excitation and temperature conditions, the model calculates the corresponding simulated terminal voltage sequence Vsim(tk;Θ) based on the parameter set Θ.

[0109] The model contains solid-phase diffusion equations, electrolyte transport equations, interface Butler–Volmer kinetics, thermal coupling equations, etc. During the identification process, attention should also be paid to its output voltage and optional intermediate states (such as local overpotential, polarization voltage, etc.).

[0110] 3.4 Objective Function Construction and Error Measurement To measure the deviation between the model output and the experimental data, this embodiment adopts an objective function based on the terminal voltage RMSE, and introduces multi-condition weighting and regularization terms to make the identification results more robust.

[0111] Voltage RMSE under a single operating condition: Assuming there are N sampling points under a certain test condition, the voltage RMSE is defined as follows: , Where, the subscript j represents the j-th operating condition, such as different rates, different temperatures, etc.; Θ is the parameter vector of the electrochemical model to be identified, including all parameters that need to be optimized, such as diffusion coefficient and exchange current density; M is the total number of operating conditions selected; N j t represents the number of voltage sampling points collected under the j-th operating condition; k represents the k-th time sampling point under this operating condition; t k V represents the time point of the kth sampling moment; exp,j (t k) represents the experimental terminal voltage measured at time t under the j-th operating condition; Vsim,j(tk;Θ) represents the simulated terminal voltage calculated by the electrochemical model with parameter Θ under the same operating condition and time; RMSE j (Θ): The root mean square error between the model voltage and the experimental voltage under the j-th operating condition, used to measure the goodness of fit of the model under that operating condition.

[0112] (2) Weighted overall objective function for multiple working conditions: If a total of M working conditions are selected, the overall objective function can be constructed as follows: , Where J(Θ): the overall objective function value obtained by integrating all selected working conditions, which is the object to be minimized in parameter identification; w j For operating condition weights, you can set equal weights based on multiplier, importance, or data quality. j =1 / M.

[0113] (3) Optional normalization error and regularization term: To avoid the influence of different dimensions or voltage ranges on the results, normalized RMSE is used or a parameter regularization term is added to the objective function, for example: , NRMSE j (Θ) represents the normalized root mean square error under the j-th working condition. Θ0 is the prior or initial parameter, Θscale is the scaling factor used to normalize parameters of different dimensions, such as taking the typical magnitude or allowable range of each parameter; λ is the regularization coefficient, which controls the weight of the regularization term in the objective function. The larger λ is, the closer the parameters tend to be to the prior value, the lower the risk of overfitting, but the fitting accuracy may decrease slightly. It is also beneficial to ensure that the parameters change within a physically reasonable range and avoid behaviors that cannot be understood by physical phenomena.

[0114] 3.5 Identification Algorithm and Optimization Process This embodiment employs an optimization framework combining global search and local refinement to solve for parameter Θ. : (1) Parameter constraints and search space setting. To ensure physical rationality, the upper and lower limits of each parameter are first determined based on material properties, literature, or experimental priors: Θ min ≤Θ≤Θ max For example, the logarithm of the diffusion coefficient. 10 D range, feasible range of electrolyte conductivity, etc.

[0115] (2) Global search stage. Traditional global optimization algorithms (such as particle swarm optimization, genetic algorithm, simulated annealing, etc.) are used to search within a limited interval so that the objective function J(Θ) converges to a smaller value region. The output of this stage is a set of candidate solutions Θ(0) that are close to the global optimum.

[0116] (3) Local refinement stage. Near the initial value Θ(0) obtained from the global search, local optimization algorithms based on gradients or quasi-Newton methods (such as Levenberg–Marquardt, BFGS, conjugate gradient, etc.) are used to finely adjust the parameters to further reduce RMSE. .

[0117] (4) Algorithm convergence criterion. The objective function decreases by less than a given threshold (e.g., ΔJ < 10). 4 The parameter update amount ||ΔΘ|| is less than the threshold; or the number of iterations reaches the set upper limit.

[0118] The above process can be implemented offline or extended to a certain extent to online / incremental updates to track the evolution trajectory of parameters during the degradation of hybrid capacitors.

[0119] 3.6 Identification Accuracy and Evaluation Indicators To evaluate the accuracy and reliability of the constructed parameter identification method, this application introduces various evaluation metrics, including but not limited to: (1) Multi-condition voltage RMSE / NRMSE. RMSE is calculated separately for each operating condition. j The maximum and average values ​​are calculated; design objectives are proposed based on requirements, such as preferably controlling the voltage RMSE under the main operating conditions within a few millivolts (e.g., within 10–20 mV) to ensure accurate fitting of the model on the SOC–voltage curve.

[0120] (2) Maximum absolute error of voltage (MaxAE): It is used to evaluate whether there are local segments with poor fit.

[0121] (3) Capacity and energy error. For a complete charge-discharge curve, the model prediction and experimental discharge capacity and energy output can be compared to calculate the relative error. .

[0122] (4) Parameter stability and repeatability. By identifying and comparing the parameter set Θ obtained from different number of cycles or repeated experimental data, The fluctuation range. If the identification results are close to each other under similar health conditions, it indicates that the algorithm is stable and robust; if it changes monotonically with the number of cycles and the trend is consistent with the degradation mechanism, it indicates that it can be used to characterize the degradation trajectory.

[0123] (5) Physical rationality verification. Check whether the obtained parameters fall within the expected physical range, such as whether the diffusion coefficient, exchange current density, porosity, etc. are within a reasonable order of magnitude, to avoid non-physical solutions caused by overfitting.

[0124] 3.7 Relationship and Application with Degradation Parameter Acquisition The parameter set Θ obtained by the above method These are the electrochemical degradation parameters corresponding to the internal mechanisms of hybrid capacitors under current aging conditions. Compared to the degradation characteristics directly measured experimentally (capacitance decay, internal resistance increase, EIS changes, etc.), the internal parameters identified through the mechanistic model in this invention are more interpretable and have physical significance. For example, the decrease in Dsn and Dsp can serve as a criterion for "diffusion-restricted" degradation modes; the continuous decrease in i0_n, i0_p, and κ indicates that interfacial kinetic degradation is dominant; the significant decrease in csn_max and csp_max corresponds to "active material loss" degradation; and the decrease in t+ reflects the weakening of electrolyte transport capacity. Therefore, the electrochemical model parameter identification method constructed in this section provides a quantifiable foundation for establishing a one-to-one mapping between mechanistic parameters and typical failure modes, as well as for realizing fault diagnosis and lifetime prediction based on parameter trajectories.

[0125] 4. A method for quantifying the fault severity of hybrid capacitors based on electrochemical model parameters In the degradation process of hybrid capacitors, multiple physical and chemical mechanisms occur simultaneously, including active material decay, SEI growth, increased charge transfer impedance, electrolyte degradation, and restricted solid-phase diffusion. Traditional single-feature methods based on capacitance or internal resistance are insufficient to comprehensively describe these mechanistic changes. Electrochemical model parameters can directly reflect internal dynamics and structural evolution; therefore, constructing a degradation quantification method based on model parameters is of great significance for interpretable fault diagnosis. This application proposes a degradation quantification framework based on model parameters, which, through data preprocessing, feature extraction, regression modeling, weight optimization, and index construction, achieves a comprehensive characterization of the multi-factor degradation of hybrid capacitors.

[0126] 4.1 Data Preparation and Preprocessing The input data for the quantification method comes from prior parameter identification and experimental measurements, including capacity, internal resistance, electrochemical impedance spectroscopy parameters (such as charge transfer resistance Rct, diffusion impedance), solid-phase diffusion coefficient Ds, porosity ε, specific surface area A, etc. Each data set is labeled with the number of cycles or degradation time, as well as actual indicators (such as capacity retention rate, internal resistance growth factor, etc.). Due to the significant differences in the dimensions of each feature, this embodiment uses the Z-score method to standardize all variables, making their mean 0 and variance 1, thereby avoiding the impact of dimensional inconsistencies on model training. Furthermore, missing data is imputed using interpolation, and noise and outliers are smoothed or removed to ensure data stability and reliability.

[0127] 4.2 Feature Selection and Feature Engineering To ensure the model's interpretability and representativeness, this application screens features strongly correlated with the degradation process based on electrochemical mechanisms. A decrease in the solid-phase diffusion coefficient typically corresponds to diffusion limitation and increased polarization, while an increase in Rct reflects deterioration in interfacial dynamics. Changes in porosity and specific surface area are closely related to microstructure shrinkage and byproduct deposition. Redundant features are further eliminated through Pearson correlation coefficient analysis to improve model stability. Simultaneously, to describe complex degradation behaviors, this embodiment constructs degradation rate-type features (such as capacity reduction rate) and cross-features (such as Rct×Ds or porosity / specific surface area) to enhance the model's ability to characterize coupled degradation mechanisms.

[0128] 4.3 Regression Modeling and Weighted Coefficient Optimization To quantitatively assess the contributions of various degradation factors, this application employs a regression model to construct a degradation degree predictor. Based on the complexity of the degradation mechanism and the correlation characteristics of features, multiple linear regression, random forest regression, and Elastic Net regression were selected for comparison. Elastic Net, which includes both L1 and L2 regularization terms, combines the sparsity of Lasso with the stability of Ridge, making it suitable for highly correlated electrochemical parameters such as diffusion coefficient, particle size, and migration number. The goal of model training is to learn the weights of each feature on the overall degradation degree, enabling it to reflect the magnitude of the contribution of different degradation factors.

[0129] During training, the model automatically adjusts its weights by optimizing the loss function (such as MSE or Elastic Net combined loss). The resulting weighted coefficients are not only used to calculate the degree of degradation, but also have a clear physical meaning: the larger the weight, the more significant the impact of the degradation mechanism corresponding to that parameter on the overall performance decline.

[0130] 4.4 Quantitative Calculation of Degradation Factors and Degree of Degradation After the model training is complete, the contribution value of each degradation factor can be calculated based on the weights learned by the model. For example, capacity decay, increased internal resistance, and decreased diffusion coefficient correspond to different degradation channels. For any parameter x_i, its degradation contribution can be expressed as: Contributioni = w_i·x_i.

[0131] Where w_i are the weighting coefficients learned by the model. Based on this, this application constructs a comprehensive degradation index to characterize the overall health status of hybrid capacitors at a certain degradation stage: Degradation Index=w1·Capacity Loss+w2·Resistance Increase+w3·Diffusion Drop+

[0132] The comprehensive degradation index can flexibly reflect the superposition effect of different mechanisms. The higher the value, the more severe the degradation of the hybrid capacitor. It can be divided into mild, moderate and severe degradation ranges, and can be further correlated with the fault type.

[0133] 4.5 Model Evaluation and Optimization To ensure the model's generalization ability and stability, this embodiment employs k-fold cross-validation to evaluate its performance under different training set partitions, and quantitatively evaluates its fitting quality using metrics such as R² and RMSE. Simultaneously, the prediction errors at different degradation stages are compared to verify the model's stability in early degradation, mid-term decay, and severe decay stages. If the model's error is too high in a certain interval, it is re-optimized by adjusting feature combinations, regularization strength, or weight distribution.

[0134] 4.6 Results Visualization and Report Output To comprehensively demonstrate the degradation mechanism and the evolution of various parameters, this application can generate heat maps, trajectory scatter plots, and comprehensive index curves to visualize the degradation trend and the contribution of degradation factors. Furthermore, based on the indices and contributions output by the model, a diagnostic report can be generated, which can be used for real-time health management and fault early warning in engineering applications.

[0135] This application constructs a method for quantifying the degradation degree of hybrid capacitors based on electrochemical model parameters. Through standardized data preprocessing, scientific feature selection, robust regression modeling, and interpretable index construction, it achieves a comprehensive evaluation of multiple degradation mechanisms. This method can not only identify degradation factors but also quantify their contribution based on model weights, providing an interpretable and scalable computational foundation for fault diagnosis, SOH assessment, and RUL prediction of hybrid capacitors.

[0136] Based on the aforementioned methods and data support, several more embodiments are given below to verify the effectiveness of the methods in the embodiments of this application.

[0137] Example 1: A fault diagnosis method for abnormal growth of SEI membrane based on model parameters This embodiment is used to verify the diagnostic capability of this application in the SEI film abnormal growth fault scenario. SEI film abnormal growth is usually accompanied by enhanced side reactions, continuous reversible lithium loss and a significant increase in interface impedance, and is a common source of early and mid-term degradation in power hybrid capacitors.

[0138] (1) Model parameter offset features Based on the aforementioned mechanism mapping relationship, when SEI grows abnormally, it exhibits the following typical parameter change characteristics: SEI thickness δ_SEI increases significantly; SEI conductivity κ_SEI decreases (ion transport deteriorates); side reaction rate constant k_SEI increases; capacity decay corresponding to reversible lithium accelerates; ohmic resistance RΩ and interfacial resistance R_film increase simultaneously. These changes will affect the parameter vector identified by the model, Θ={Dsn,Dsp,cs_max,i0,k SEI ,κ SEI Significant offset is observed in ,…}.

[0139] (2) Fault feature extraction Using the constructed model parameter identification method (RMSE optimization strategy), parameter inversion is performed for each loop node, and the following residual features are constructed, z_k SEI : Standardized shift of side reaction rates; z_κ SEI SEI conductivity relative decrease; Δcapacity_slope: capacity decrease slope over the last 100 cycles; R film_residual The residual increase of R_film. The above characteristics constitute the fault characteristics.

[0140] (3) Diagnostic rules The criteria for determining abnormal growth of the SEI membrane are defined as follows: , Where α1–α4 are thresholds obtained statistically from a normal degradation database. Once any 2 to 3 conditions are met simultaneously, it is determined that the SEI membrane is growing abnormally.

[0141] (4) Quantification of Fault Severity Using the SEI composite index: SEI_Index=w1zk SEI +w2zκ SEI +w3R film_residual+w4Δcapacity_slope Where wi is automatically obtained from regression training, and the final exponent is normalized to 0–1, corresponding to mild (0–0.3), moderate (0.3–0.6), and severe (>0.6).

[0142] (5) Applicability of the project This embodiment is applicable to hybrid capacitors that experience sudden capacity loss after long-term static storage, SEI layer regeneration phenomenon in fast-charging hybrid capacitors, and regeneration cycle scenarios after SEI rupture under low-temperature conditions.

[0143] Example 2: A Model Parameter-Based Method for Diagnosing Loss-of-Active-Material (LAM) Faults This embodiment is used to verify the diagnostic capability of this application under the Loss of Active Material (LAM) fault type. LAM typically originates from particle breakage, pulverization, detachment, or electrode structure collapse.

[0144] (1) Model parameter offset features Typical parameter changes include: continuous decrease in maximum solid concentrations csmax_p and csmax_n at the positive and negative electrodes; significant decrease in solid volume fraction εs; significant decrease in exchange current density i0 (reduction in effective specific surface area A); decrease in solid conductivity σs (disruption of electron pathways); and a significant acceleration in capacity decay rate.

[0145] (2) Fault feature construction The following features were obtained through model identification: z_csmax_p, z_csmax_n: maximum decrease in solid-phase lithium concentration; z_eps: decrease in electrode porosity; z_i0: decrease in exchange current density; and capacity decay slope dQ / dN. These features strongly indicate LAM failure.

[0146] (3) Diagnostic criteria A hybrid capacitor is considered to have entered a LAM fault when the following conditions are met: , A fault can be identified as being dominated by the loss of active materials if any two of the conditions are met.

[0147] (4) Fault severity calculation Constructing the LAM index: LAM_Index=v1z cs _ max +v2z i0 +v3z εs +v4∣dQ / dN∣。 The index value is mapped to the mild, moderate, and severe LAM range.

[0148] (5) Engineering application scenarios: mid-to-late cycle life; particle fracture caused by high stress / high rate; structural failure diagnosis of high nickel cathode materials.

[0149] Example 3: Electrolyte Deterioration Fault Diagnosis Method Based on Model Parameters This embodiment applies to electrolyte degradation scenarios, including electrolyte decomposition, salt depletion, viscosity increase, and solvent decomposition product deposition.

[0150] (1) Model parameter offset features Electrolyte degradation corresponds to changes in the following mechanistic parameters: a significant decrease in electrolyte conductivity κ; a decrease in electrolyte diffusion coefficient De; an abnormal shift in transport number t+; accelerated growth in DC internal resistance (DCIR); and an increase in high-frequency impedance Rs. (2) Feature extraction Through the identification model, z_κ is obtained as follows: electrolyte conductivity residual; z_De is the diffusion coefficient offset; z_tplus is the transport number offset; DCIR_slope is... DCIR / cycle; the combination of characteristics forms the electrolyte degradation characteristics.

[0151] (3) Fault Judgment Rules , If any two of the following conditions are met, the electrolyte is considered to be deteriorated.

[0152] (4) Fault severity rating ED_Index=u1z κ +u2z De +u3z t+ +u4(DCIR slope) This allows for differentiation of the degree of electrolyte malfunction.

[0153] (5) Application scenarios: high temperature cycling, high rate charging and discharging, electrolyte oxidation / decomposition risk, long-term float charging application of energy storage system.

[0154] Example 4: A Multi-Mechanism Synergistic Degradation Fault Diagnosis Method This embodiment is used to diagnose the multiple degradation coupled fault, i.e., the simultaneous occurrence of multiple mechanisms such as SEI growth, active material loss, and electrolyte degradation. This situation is most common in real hybrid capacitors (such as under high temperature, high rate, and long-term aging conditions).

[0155] (1) Parameter offset characteristics This manifests as: simultaneous shifts in multiple parameters, lacking a single dominant mechanism; nonlinear and abrupt changes in parameter trends; and deterioration in capacity, internal resistance, and polarization. Specifically, the following will occur simultaneously: increased k_SEI, decreased κ_SEI (interfacial degradation), decreased csmax_p / n and εs (loss of active material), decreased κ and De (electrolyte anomalies), and decreased Ds and i0 (diffusion and kinetic degradation). (2) Feature combination Construct the joint feature vector: F=[zk SEI ,z_κ SEI ,z cs,max ,z εs ,z κ ,z De ,z Ds ,z i0 The compositional relationships are learned through Elastic-Net regression.

[0156] (3) Diagnostic methods Define the hybrid residual metric: Hybrid Residual = ||F||², and set a threshold δ: ||F||² > δ Multi-mechanism synergistic degradation. When multiple mechanistic indicators exceed the normal range of a single mechanism, but do not meet the fault criteria of any single mechanism, it is judged as multi-mechanism degradation.

[0157] (4) Failure severity index A comprehensive weighted model is adopted: MDI=η1SEI_Index+η2LAM_Index+η3ED_Index, which is then normalized to 0–1 to give the severity of collaborative faults.

[0158] (5) Engineering significance This embodiment is particularly suitable for: real vehicle operating environments, high temperature and high rate coupling, multi-mechanism coupling degradation caused by fast charging, and strong degradation regions at the end of the entire life cycle.

[0159] Based on the description of the foregoing method embodiments, this application also provides a fault diagnosis device based on a hybrid capacitor electrochemical model.

[0160] like Figure 3 As shown, the fault diagnosis device 300 based on a hybrid capacitor electrochemical model includes: The data acquisition module 310 is used to acquire experimental data of the hybrid capacitor at multiple degradation stages, including capacitance, internal resistance, electrochemical impedance spectroscopy, and voltage-time curves. The model building module 320 is used to establish an electrochemical model that includes solid-phase diffusion, interfacial reaction and electrolyte transport, and selects multiple mechanism parameters that are sensitive to degradation as a set to be identified. The parameter identification module 330 is used to identify the parameters of the set to be identified based on the experimental voltage curve and with the terminal voltage error as the objective function, so as to obtain the mechanism parameter sequence that evolves with the degradation stage. Feature extraction module 340 is used to extract the offset features of the mechanism parameter sequence; The fault mapping module 350 is used to establish a mapping relationship between the offset features and preset typical fault modes in order to identify the fault type; The fault quantification module 360 ​​is used to construct a weighted degradation index based on the offset features to quantify the severity of the fault.

[0161] It is understood that the relevant content concerning each module in the above-mentioned device has been described in detail in the foregoing method embodiments, and specific details can be found in the method embodiments; that is, the fault diagnosis device 300 based on a hybrid capacitor electrochemical model provided in this application can perform the following... Figure 1 Any steps in the illustrated embodiments will not be described in detail here.

[0162] In one embodiment of this application, an electronic device is also provided. See also... Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4 As shown, the electronic device 400 includes a processor 401 and a memory 402. The memory 402 stores a computer program, which, when executed by the processor 401, will perform actions such as... Figure 1 Any step in the method embodiment shown can be a control method step in the experimental process, such as controlling and adjusting the transformer ratio of the voltage regulator, monitoring the phase difference, etc. The electronic device 400 may also include input / output devices, etc. In a specific embodiment, the electronic device can be a terminal device, etc.

[0163] In one embodiment, a computer-readable storage medium is also provided, which stores a computer program that, when executed by a processor 401, causes the processor 401 to perform any of the steps in the above method embodiments.

[0164] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0165] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0166] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A fault diagnosis method based on a hybrid capacitor electrochemical model, characterized by, The method comprises: obtaining experimental data of the hybrid capacitor at multiple degradation stages, the experimental data comprising capacity, internal resistance, electrochemical impedance spectrum, and voltage-time curve; establishing an electrochemical model comprising solid-phase diffusion, interface reaction, and electrolyte transmission, and selecting multiple mechanism parameters sensitive to degradation as a to-be-identified set; based on the experimental voltage curve, adopting an optimization strategy taking terminal voltage error as an objective function to perform parameter identification on the to-be-identified set, and obtaining a mechanism parameter sequence evolving with the degradation stage; extracting a shift feature of the mechanism parameter sequence; establishing a mapping relationship between the shift feature and a preset typical fault mode to identify a fault type; constructing a weighted degradation index based on the shift feature to quantify a fault severity.

2. The method of claim 1, wherein the method is based on a hybrid capacitor electrochemical model. The electrochemical model further comprises an SEI growth process. The to-be-identified set at least comprises a solid-phase diffusion coefficient, a maximum solid-phase concentration, porosity, an electrolyte diffusion coefficient, an exchange current density, a side reaction rate constant, an SEI thickness, or an equivalent impedance parameter thereof.

3. The method of claim 1, wherein the method is characterized by: In the electrochemical model, the solid-phase diffusion coefficient, the maximum solid-phase concentration, the active material volume fraction, the electrode porosity, the exchange current density, the charge transfer rate constant, the electrolyte conductivity, the electrolyte diffusion coefficient, the transference number, the SEI film thickness, the conductivity, and the side reaction rate constant in the to-be-identified set are respectively mapped to a hybrid capacitor degradation fault label one by one to form an intermediate basis for fault diagnosis.

4. The method of claim 1, wherein the method is based on a hybrid capacitor electrochemical model. The optimization strategy adopts a framework combining global search and local refinement, and is jointly executed under multiple working condition experimental data.

5. The method of claim 1, wherein the method is based on a hybrid capacitor electrochemical model. The shift feature comprises a parameter change rate, a standardized residual, and a cross-parameter coupling feature for characterizing multiple mechanism collaborative degradation.

6. The method of claim 1, wherein the method is based on a hybrid capacitor electrochemical model. The typical fault mode comprises SEI abnormal growth, active material loss, electrolyte degradation, and a combination thereof.

7. The method of claim 1, wherein the method further comprises: The fault type comprises: SEI abnormal growth; positive or negative active material loss; electrolyte degradation; diffusion-limited degradation; pore blockage type degradation; multiple mechanism collaborative degradation.

8. The method of claim 1, wherein the method is based on a hybrid capacitor electrochemical model. The obtaining of the experimental data of the hybrid capacitor at multiple degradation stages comprises: testing performance parameters of the hybrid capacitor at multiple cycles, the performance parameters comprising capacity, coulomb efficiency, internal resistance, charge transfer resistance, diffusion coefficient, heat generation rate, thermal conductivity / specific heat; corresponding the fault indicators corresponding to the performance parameters with the cycle number to construct a degradation fault data set of the hybrid capacitor.

9. A fault diagnosis apparatus based on a hybrid capacitor electrochemical model, characterized by, It comprises: a data acquisition module configured to obtain experimental data of a hybrid capacitor at multiple degradation stages, the experimental data comprising capacity, internal resistance, electrochemical impedance spectrum, and voltage-time curve; a model construction module configured to establish an electrochemical model comprising solid-phase diffusion, interface reaction, and electrolyte transmission, and select multiple mechanism parameters sensitive to degradation as a to-be-identified set; a parameter identification module configured to, based on an experimental voltage curve, adopt an optimization strategy taking terminal voltage error as an objective function to perform parameter identification on the to-be-identified set, and obtain a mechanism parameter sequence evolving with the degradation stage; a feature extraction module configured to extract a shift feature of the mechanism parameter sequence; A fault mapping module is configured to establish a mapping relationship between the offset feature and a preset typical fault mode, so as to identify a fault type. A fault quantification module is configured to construct a weighted degradation index based on the offset feature, and to quantify a fault severity.

10. An electronic device, comprising: A computer program product is provided, comprising a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the method according to any one of claims 1-7.

Images