Nuclear-grade lead-acid battery fault prediction method, system and storage medium

Abstract

The present invention discloses a nuclear-grade lead-acid battery fault prediction method, system and storage medium, the method comprising: S10, obtaining sample aging data during the aging process of the lead-acid battery; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure and property change characteristics; S20, establishing a battery equivalent circuit model based on the internal structure and property change characteristics; S30, performing multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establishing a multidimensional mapping model based on the multidimensional feature data and the available capacity data; the multidimensional mapping model is used to cooperate with the battery equivalent circuit model to predict the degradation process of the battery's available capacity; S40, using the battery data to be tested to train the battery equivalent circuit model, and predicting the degradation of the available capacity of the battery to be tested based on the training results; implementation of the present invention can accurately predict the available capacity of the battery to be tested.

Description

Technical Field

[0001] The present invention relates to the technical field of lead-acid batteries, and in particular to a nuclear-grade lead-acid battery fault prediction method, system and storage medium. Background Art

[0002] Nuclear-grade lead-acid batteries in nuclear power plants are complex electrochemical systems with sealed structures. In practical applications, changes in capacity or internal resistance are often used as indicators of battery health. For backup power supplies, users are more concerned with the battery's ability to discharge continuously in an emergency. Therefore, battery management systems typically simplify battery health diagnosis to simply estimating available capacity.

[0003] Currently, the estimation of the available capacity of lead-acid batteries is mainly divided into offline capacity testing and online estimation. Offline capacity testing involves disconnecting the battery pack from the power supply system and then performing a discharge test to obtain the remaining capacity of the battery pack. This method is time-consuming and cumbersome to operate. When performing offline capacity testing, the stability of the power supply system cannot be guaranteed. Online estimation requires modeling the lead-acid battery and constructing an available capacitance evaluation model. The available capacitance evaluation model is then input into the available capacitance evaluation model using the online detected lead-acid battery voltage, current, and other parameters to obtain the available capacitance of the lead-acid battery. However, existing evaluation models generally have poor prediction accuracy and may even require the modification of on-site equipment, which increases the measurement workload for on-site operation and maintenance personnel, making it difficult to achieve large-scale application in industrial sites (nuclear power plants). Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a nuclear-grade lead-acid battery fault prediction method, system and storage medium.

[0005] The technical solution adopted by the present invention to solve the technical problem is to construct a nuclear-grade lead-acid battery fault prediction method, comprising the following steps:

[0006] S10, obtaining sample aging data of the lead-acid battery during the aging process; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure and property change characteristics;

[0007] S20, establishing a battery equivalent circuit model based on the internal structure and property change characteristics;

[0008] S30, performing multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establishing a multidimensional mapping model based on the multidimensional feature data and the available capacity data; the multidimensional mapping model is used to cooperate with the battery equivalent circuit model to predict the degradation process of the battery's available capacity;

[0009] S40: Using the test battery data to train the battery equivalent circuit model, and predicting the degradation of the available capacity of the test battery according to the training result.

[0010] Preferably, in S20, the battery equivalent circuit model includes a first inductor Lp, a second inductor Ln, a first capacitor C1, a second capacitor C2, a first charge transfer resistor R1, a second charge transfer resistor R2, a first impedance Z1, a second impedance Z2, an ohmic resistor Rs, a floating charge power source, and a virtual current source;

[0011] One path of the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first capacitor C1, another path of the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first charge transfer resistor R1 and the first impedance Z1, one path of the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second capacitor C2, another path of the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second charge transfer resistor R2 and the second impedance Z2, the first end of the first inductor Lp is connected to the virtual current source, the virtual current source is connected to the positive electrode of the floating charge power supply, and the negative electrode of the floating charge power supply is connected to the second end of the second inductor Ln;

[0012] The multi-dimensional mapping model is connected to the virtual current source to control the output of the virtual current source.

[0013] Preferably, the nuclear-grade lead-acid battery fault prediction method further includes:

[0014] S31 . Establish an indirect mapping relationship between the battery equivalent circuit model and the multidimensional mapping model, so as to set input parameters of the multidimensional mapping model according to the indirect mapping relationship and design parameters of the battery equivalent circuit model.

[0015] Preferably, in S30, the multidimensional feature extraction process includes:

[0016] extracting multidimensional statistical features from the float charge voltage data using a descriptive statistical method, clustering the multidimensional statistical features using a clustering algorithm, and determining a multidimensional mapping relationship between the multidimensional statistical features and the available capacity data based on the clustering results;

[0017] The multidimensional feature data includes multidimensional statistical features and multidimensional mapping relationships.

[0018] Preferably, the extracting of multidimensional statistical features from the float charge voltage data using a descriptive statistical method includes:

[0019] By analyzing the spectrum, frequency and information entropy of the float charge voltage data, a one-dimensional feature in the multidimensional statistical feature is obtained; and discrete statistics are performed on the one-dimensional feature based on time and different sample batteries to obtain a multidimensional feature of not less than two dimensions in the multidimensional statistical feature.

[0020] Preferably, the step of extracting multidimensional statistical features from the float charge voltage data using a descriptive statistical method further includes:

[0021] According to the preset electrode potential change law of the battery degradation process, the type of the descriptive statistical method to be used when extracting different features is selected.

[0022] Preferably, the clustering algorithm includes at least one of an isolation forest algorithm, a t-SNE algorithm, and fold cross validation.

[0023] Preferably, in said S10, said property change characteristics include at least one of electrode corrosion change data, electrolyte change data, active material change data, passivation layer growth data and sulfate crystallization data.

[0024] The present invention also constructs a storage medium storing a computer program, wherein the computer program instructions, when executed by a processor, implement the steps of the nuclear-grade lead-acid battery fault prediction method provided in an embodiment of the present invention.

[0025] The present invention also constructs a nuclear-grade lead-acid battery fault prediction system, which includes:

[0026] An acquisition unit, configured to acquire sample aging data of a lead-acid battery during its aging process; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure, and property change characteristics;

[0027] A first modeling unit is used to establish a battery equivalent circuit model according to the internal structure and property change characteristics;

[0028] a second modeling unit, configured to perform multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establish, based on the multidimensional feature data and the available capacity data, a multidimensional mapping model for cooperating with the battery equivalent circuit model to predict a degradation process of the available capacity of the battery;

[0029] A training unit is used to train the battery equivalent circuit model using the battery data to be tested, and predict the degradation of the available capacity of the battery to be tested based on the training results.

[0030] To implement the technical solution of the present invention, first, a battery equivalent circuit model is established based on the internal structure and property change characteristics; then, multidimensional feature extraction processing is performed on the float charge voltage data to obtain multidimensional feature data, and a multidimensional mapping model is established based on the multidimensional feature data and available capacity data. The multidimensional mapping model is then coordinated with the battery equivalent circuit model to improve the prediction accuracy of the battery equivalent circuit model; then, the battery equivalent circuit model is trained using the battery data to enhance the usability and robustness of the battery equivalent circuit model in an actual environment, and the available capacity of the battery to be tested is accurately predicted based on the training results. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0032] Figure 1 is a flow chart of a method for predicting nuclear-grade lead-acid battery failure in some embodiments of the present invention;

[0033] Figure 2 is a circuit structure diagram of a battery equivalent circuit model in some embodiments of the present invention;

[0034] Figure 3 Schematic diagram of the structure of a nuclear-grade lead-acid battery fault prediction system in some embodiments of the present invention. DETAILED DESCRIPTION

[0035] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, specific embodiments of the present invention are now described in detail with reference to the accompanying drawings.

[0036] It should be noted that the flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all content and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0037] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0038] See also Figure 1 FIG2 is a flow chart of a method for predicting nuclear-grade lead-acid battery failures in some embodiments of the present invention. The method is applied to a host computer to predict the available capacity of a nuclear-grade lead-acid battery. The method includes the following steps:

[0039] S10, obtaining sample aging data of the lead-acid battery during the aging process; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure and property change characteristics;

[0040] S20. Establishing a battery equivalent circuit model based on the internal structure and property change characteristics;

[0041] S30, performing multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establishing a multidimensional mapping model based on the multidimensional feature data and the available capacity data; the multidimensional mapping model is used to cooperate with the battery equivalent circuit model to predict the degradation process of the battery's available capacity;

[0042] S40: Using the data of the battery to be tested to train a battery equivalent circuit model, and predicting degradation of the available capacity of the battery to be tested based on the training results.

[0043] The sample aging data may include aging data recorded during the aging process of multiple nuclear-grade lead-acid batteries. Accordingly, the float charge voltage data includes recorded data on the change in float charge voltage over time for each nuclear-grade lead-acid battery during the aging process; the available capacity data includes recorded data on the change in available capacity over time for each nuclear-grade lead-acid battery during the aging process; the internal structure includes the structural characteristics of each component within the nuclear-grade lead-acid battery, such as the connection structure between components such as the positive electrode, negative electrode, and electrolyte; and the property change characteristics include at least one of electrode corrosion change data, electrolyte change data, active material change data, passivation layer growth data, and sulfate crystallization data for each nuclear-grade lead-acid battery during the aging process. Furthermore, the internal structure and property change characteristics can be determined by measuring the dynamic changes in the physical and chemical properties of the battery electrode deposit layer, such as the micromorphology and component content, at different levels of aging.

[0044] For the battery equivalent circuit model, the equivalent electronic components of each component inside the nuclear-grade lead-acid battery, as well as the topological structure of these equivalent electronic components, are determined through the internal structure and property change characteristics. In this way, a battery equivalent circuit model for quantitative analysis of the battery degradation mechanism is established. The battery equivalent circuit model is trained using the battery data to calculate the coupling relationship between the internal structure and property change characteristics of the battery to be tested and the available capacity data, making battery aging explainable at the reaction kinetics level, which is conducive to improving the accuracy of battery available capacity prediction.

[0045] It should be noted that in the late stage of degradation, the float charge voltage and available capacity of lead-acid batteries will drop sharply, and the internal resistance will rise sharply. Moreover, there is a clear order in the conversion moments of the sharp changes. Usually, the conversion moment of available capacity appears first, followed by the internal resistance, and the float charge voltage appears last. Therefore, by performing multi-dimensional feature extraction processing on the float charge voltage data, the closed-loop feedback effect between the float charge voltage fluctuation and the evolution of the degradation process can be explored, and the multi-dimensional feature data of the float charge voltage during the aging process can be determined. In this way, the multi-dimensional feature data can be used to predict the available capacity and remaining life of the battery. Therefore, a multi-dimensional mapping model can be established based on the multi-dimensional feature data and available capacity data. By inputting the relevant parameters of the battery to be tested into the multi-dimensional mapping model, the mapping relationship between the multi-dimensional feature data and the available capacity data of the battery to be tested can be obtained, and then combined with the battery equivalent circuit model to predict the degradation process of the battery's available capacity.

[0046] This embodiment first establishes a battery equivalent circuit model based on the internal structure and property change characteristics; then performs multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establishes a multidimensional mapping model based on the multidimensional feature data and available capacity data. The multidimensional mapping model is then coordinated with the battery equivalent circuit model to improve the prediction accuracy of the battery equivalent circuit model; then, the battery equivalent circuit model is trained using the battery data to enhance the usability and robustness of the battery equivalent circuit model in an actual environment, and the available capacity of the battery to be tested is accurately predicted based on the training results.

[0047] In an alternative embodiment, if Figure 2 As shown, the battery equivalent circuit model includes a first inductor Lp, a second inductor Ln, a first capacitor C1, a second capacitor C2, a first charge transfer resistor R1, a second charge transfer resistor R2, a first impedance Z1, a second impedance Z2, an ohmic resistor Rs, a floating power supply 80, and a virtual current source. One path from the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first capacitor C1. Another path from the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first charge transfer resistor R1 and the first impedance Z1. One path from the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second capacitor C2. Another path from the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second charge transfer resistor R2 and the second impedance Z2. The first end of the first inductor Lp is connected to the virtual current source, which is connected to the positive terminal of the floating power supply 80, and the negative terminal of the floating power supply 80 is connected to the second end of the second inductor Ln. The multidimensional mapping model is connected to the virtual current source to control the output of the virtual current source.

[0048] Among them, the first inductor Lp represents the equivalent inductance of the positive electrode; the second inductor Ln represents the equivalent inductance of the negative electrode; the first capacitor C1 represents the equivalent capacitance caused by the spatial charge distribution of the electrochemical double layer on the positive electrode; the second capacitor C2 represents the equivalent capacitance caused by the spatial charge distribution of the electrochemical double layer on the negative electrode; the first charge transfer resistor R1 represents the equivalent charge transfer resistance on the positive electrode; the second charge transfer resistor R2 represents the equivalent charge transfer resistance on the negative electrode; the first impedance Z1 represents the equivalent diffusion impedance (Warburg impedance) caused by the diffusion of ions on the positive electrode in the electrolyte and electrode pores; the second impedance Z2 represents the equivalent diffusion impedance caused by the diffusion of ions on the negative electrode in the electrolyte and electrode pores; the ohmic resistance Rs represents the ohmic resistance caused by various internal structures of the battery (such as connection resistance, separator resistance, electrolyte resistance and surface coverage of crystalline lead sulfate electrode, etc.). In this embodiment, the battery data to be tested includes the float charge voltage data, available capacity data, internal structure, property change characteristics, battery voltage and battery internal resistance of the battery to be tested. Therefore, the battery data to be tested can be used to train the battery equivalent circuit model to determine the parameter values ​​of the first inductor Lp, the second inductor Ln, the first capacitor C1, the second capacitor C2, the first charge transfer resistor R1, the second charge transfer resistor R2, the first impedance Z1, the second impedance Z2, the ohmic resistor Rs and the float charge power source 80.

[0049] Since the electrode potential fluctuations generated by the electrochemical reactions inside the battery have a superposition effect, in order to quantitatively describe the interaction between the float charge voltage fluctuations and the battery degradation process, the parameter values ​​of the virtual current source are controlled through a multidimensional mapping model to simulate the aging of the battery. In this way, the battery equivalent circuit model can take into account the accuracy of the physical model and the dynamic characteristics of the abstract model, and then realize the quantitative study of the dynamic change mechanism of the battery under test to improve the accuracy of the battery available capacity prediction.

[0050] In an optional embodiment, the nuclear-grade lead-acid battery fault prediction method also includes: S31, establishing an indirect mapping relationship between the battery equivalent circuit model and the multidimensional mapping model, so as to set the input parameters of the multidimensional mapping model according to the indirect mapping relationship and the design parameters of the battery equivalent circuit model.

[0051] like Figure 2As shown, in this embodiment, the input parameters of the multidimensional mapping model are a virtual feedback parameter set, and the virtual feedback parameter set includes parameter values ​​of the first capacitor C1, the second capacitor C2, the first charge transfer resistor R1, the second charge transfer resistor R2, the first impedance Z1, the second impedance Z2, and the ohmic resistor Rs. The indirect mapping relationship is established so that after the battery equivalent circuit model training is completed and the parameter values ​​of the virtual feedback parameter set are determined, they are input into the multidimensional mapping model. The multidimensional mapping model thus controls the virtual current source, forming a mutual feedback mechanism, thereby further improving the prediction accuracy.

[0052] In an optional embodiment, the multidimensional feature extraction processing includes: using descriptive statistical methods to extract multidimensional statistical features from the float charge voltage data, and clustering the multidimensional statistical features through a clustering algorithm, so as to determine the multidimensional mapping relationship between the multidimensional statistical features and the available capacity data in combination with the clustering results; the multidimensional feature data includes multidimensional statistical features and multidimensional mapping relationships.

[0053] Float charge voltage fluctuations include changes in electrode potential caused by different degradation characteristics. Therefore, float charge voltage fluctuations are typically the result of the random superposition of multiple electrode potential changes. Different degradation characteristics give rise to different component characteristics in electrode potential changes. Therefore, appropriate descriptive statistical methods can be used to extract multidimensional statistical features corresponding to different degradation characteristics from float charge voltage data.

[0054] In an optional embodiment, the step of extracting multidimensional statistical features from float charge voltage data using a descriptive statistical method includes: obtaining one-dimensional features in the multidimensional statistical features by analyzing the spectrum, frequency and information entropy of the float charge voltage data; and performing discrete statistics on the one-dimensional features based on time and different sample batteries to obtain multidimensional features of no less than two dimensions in the multidimensional statistical features.

[0055] In an optional embodiment, the step of extracting multidimensional statistical features from float charge voltage data using a descriptive statistical method further includes: selecting the type of descriptive statistical method to be used when extracting different features based on a preset law of electrode potential change during the battery degradation process.

[0056] It should be noted that the electrode potential change characteristics caused by degradation processes such as electrode corrosion, electrolyte changes, active material loss, passivation layer growth, and sulfate crystallization are all different. Therefore, when the battery is in the floating charge state, multiple degradation processes usually occur simultaneously, and the random superposition of multiple degradation processes will lead to random destructive or constructive changes in the electrode potential. This makes it difficult to establish a relationship between the change in the floating charge voltage and a specific degradation feature. In this embodiment, corresponding descriptive statistical methods are designed for different degradation features to achieve a quantitative description of the electrode potential change during the battery degradation process. For example, changes in electrolyte concentration will produce a random sequence of at least one charge pulse, which will lead to fluctuations in the floating charge voltage. As the loss of electrolyte increases, the fluctuations will become more frequent and disordered. Therefore, the change in electrolyte concentration can be described by a one-dimensional statistical feature. In this embodiment, the information entropy in the float charge voltage is used to describe this disordered state. For example, electrode plate corrosion can lead to an increase in the peak-to-peak value range of the electrode potential, but this increase in peak-to-peak value does not increase the disorder of the float charge voltage. Therefore, information entropy cannot effectively characterize such degradation characteristics. This embodiment extracts such degradation characteristics through a two-dimensional discrete distribution statistical method composed of frequency and time in the float charge voltage. The statistical method is as follows: construct a coordinate system with time as the horizontal axis and the float charge voltage value as the vertical axis, set a sampling point on the horizontal axis of the coordinate system with a set step size (which can be 10,000 seconds), and set a sampling point on the vertical axis for each set interval (which can be 0.02V), forming several sampling interval blocks, and then counting the frequency of statistical values ​​falling into each interval block to describe the battery degradation characteristics caused by electrode plate corrosion. It is easy to understand that the use of descriptive statistical methods is an effective way to extract component features from the random superposition of float charge voltage fluctuations, and can effectively establish a mapping relationship between float charge voltage and specific degradation characteristics.

[0057] In an optional embodiment, in the step of extracting multidimensional statistical features from the float charge voltage data using a descriptive statistical method, it also includes: in the process of extracting the multidimensional statistical features, if the number of samples of the float charge voltage data is greater than a set number, the float charge voltage data is filtered out by the superposition method to simplify the extraction process and improve the extraction efficiency.

[0058] In an optional embodiment, the clustering algorithm includes at least one of an isolation forest algorithm, a t-SNE algorithm, and fold cross validation.

[0059] It is understandable that due to the increase in the number of nuclear-grade lead-acid batteries that may be involved in the sample aging data and the increase in the total amount of data of multidimensional statistical features, in order to save the amount of calculation related to the multidimensional statistical features, this embodiment clusters the multidimensional statistical features through a clustering algorithm, which not only helps to save the amount of calculation, but also improves the feature extraction effect of the multidimensional statistical features, and plays a positive role in improving the accuracy of battery available capacity prediction.

[0060] The present invention also provides a storage medium storing a computer program, wherein when the computer program instructions are executed by a processor, the steps of the nuclear-grade lead-acid battery fault prediction method provided in an embodiment of the present invention are implemented.

[0061] See also Figure 3 , which is a structural diagram of a nuclear-grade lead-acid battery fault prediction system in some embodiments of the present invention. The nuclear-grade lead-acid battery fault prediction system includes an acquisition unit, a first modeling unit, a second modeling unit, and a training unit.

[0062] The acquisition unit is used to acquire sample aging data during the aging process of the lead-acid battery; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure and property change characteristics.

[0063] The first modeling unit is used to establish a battery equivalent circuit model based on the internal structure and property change characteristics.

[0064] The second modeling unit is used to perform multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and to establish a multidimensional mapping model based on the multidimensional feature data and the available capacity data for use in conjunction with the battery equivalent circuit model to predict the degradation process of the battery's available capacity.

[0065] The training unit is used to train a battery equivalent circuit model using the battery data to be tested, and to predict the degradation of the available capacity of the battery to be tested based on the training results.

[0066] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0067] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the components and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0068] It is understandable that the above embodiments only express the preferred implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the patent scope of the present invention. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present invention, the above technical features can be freely combined, and several deformations and improvements can be made, all of which fall within the scope of protection of the present invention. Therefore, all equivalent changes and modifications made to the scope of the claims of the present invention should fall within the scope of coverage of the claims of the present invention.

Claims

1. A method for predicting nuclear-grade lead-acid battery failure, characterized in that: The following steps are involved: S10, obtaining sample aging data of the lead-acid battery during the aging process; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure and property change characteristics; S20, establishing a battery equivalent circuit model based on the internal structure and property change characteristics; S30: Performing multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establishing a multidimensional mapping model based on the multidimensional feature data and the available capacity data; the multidimensional mapping model is used to cooperate with the battery equivalent circuit model to predict the degradation process of the available capacity of the battery; and further comprising: S31: Establishing an indirect mapping relationship between the battery equivalent circuit model and the multidimensional mapping model, so as to set input parameters of the multidimensional mapping model based on the indirect mapping relationship and design parameters of the battery equivalent circuit model; S40, training the battery equivalent circuit model using the test battery data, and predicting the degradation of the available capacity of the test battery according to the training results; The battery equivalent circuit model includes a first inductor Lp, a second inductor Ln, a first capacitor C1, a second capacitor C2, a first charge transfer resistor R1, a second charge transfer resistor R2, a first impedance Z1, a second impedance Z2, an ohmic resistor Rs, a floating charge power source (80) and a virtual current source; one path of the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first capacitor C1, and the other path of the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first charge transfer resistor R1 and the first impedance Z1, and the ohmic resistor Rs is connected to the first end of the ohmic resistor Rs. One path of the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second capacitor C2, and the other path of the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second charge transfer resistor R2 and the second impedance Z2. The first end of the first inductor Lp is connected to the virtual current source, the virtual current source is connected to the positive electrode of the floating charge power supply (80), and the negative electrode of the floating charge power supply (80) is connected to the second end of the second inductor Ln; the multi-dimensional mapping model is connected to the virtual current source to control the output of the virtual current source.

2. The nuclear-grade lead-acid battery fault prediction method according to claim 1, characterized in that: In the S30, the multi-dimensional feature extraction process includes: extracting multidimensional statistical features from the float charge voltage data using a descriptive statistical method, clustering the multidimensional statistical features using a clustering algorithm, and determining a multidimensional mapping relationship between the multidimensional statistical features and the available capacity data based on the clustering results; The multidimensional feature data includes multidimensional statistical features and multidimensional mapping relationships.

3. The nuclear-grade lead-acid battery fault prediction method according to claim 2, characterized in that: The extracting of multidimensional statistical features from the float charge voltage data using a descriptive statistical method includes: By analyzing the spectrum, frequency and information entropy of the float charge voltage data, a one-dimensional feature in the multidimensional statistical feature is obtained; and discrete statistics are performed on the one-dimensional feature based on time and different sample batteries to obtain a multidimensional feature of not less than two dimensions in the multidimensional statistical feature.

4. The nuclear-grade lead-acid battery fault prediction method according to claim 3, characterized in that: The step of extracting multidimensional statistical features from the float charge voltage data using a descriptive statistical method further includes: According to the preset electrode potential change law of the battery degradation process, the type of the descriptive statistical method to be used when extracting different features is selected.

5. The nuclear-grade lead-acid battery fault prediction method according to claim 2, characterized in that: The clustering algorithm includes at least one of an isolation forest algorithm, a t-SNE algorithm, and fold cross validation.

6. The nuclear-grade lead-acid battery fault prediction method according to claim 1, characterized in that: In the step S10 , the property change characteristics include at least one of electrode corrosion change data, electrolyte change data, active material change data, passivation layer growth data, and sulfate crystallization data.

7. A storage medium storing a computer program, characterized in that: When the computer program instructions are executed by a processor, the steps of the nuclear-grade lead-acid battery fault prediction method according to any one of claims 1 to 6 are implemented.

8. A nuclear-grade lead-acid battery fault prediction system, characterized in that: include: An acquisition unit, configured to acquire sample aging data of a lead-acid battery during its aging process; wherein the sample aging data includes float charge voltage data, available capacity data, internal structure, and property change characteristics; A first modeling unit is used to establish a battery equivalent circuit model according to the internal structure and property change characteristics; a second modeling unit, configured to perform multidimensional feature extraction processing on the float charge voltage data to obtain multidimensional feature data, and establish, based on the multidimensional feature data and the available capacity data, a multidimensional mapping model for cooperating with the battery equivalent circuit model to predict the degradation process of the available capacity of the battery, the model comprising: establishing an indirect mapping relationship between the battery equivalent circuit model and the multidimensional mapping model, and setting input parameters of the multidimensional mapping model based on the indirect mapping relationship and design parameters of the battery equivalent circuit model; A training unit, configured to train the battery equivalent circuit model using the battery data to be tested, and predict the degradation of the available capacity of the battery to be tested based on the training results; The battery equivalent circuit model includes a first inductor Lp, a second inductor Ln, a first capacitor C1, a second capacitor C2, a first charge transfer resistor R1, a second charge transfer resistor R2, a first impedance Z1, a second impedance Z2, an ohmic resistor Rs, a floating charge power supply (80) and a virtual current source; one path of the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first capacitor C1, another path of the second end of the first inductor Lp is connected to the first end of the ohmic resistor Rs via the first charge transfer resistor R1 and the first impedance Z1, one path of the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second capacitor C2, another path of the second end of the ohmic resistor Rs is connected to the first end of the second inductor Ln via the second charge transfer resistor R2 and the second impedance Z2, the first end of the first inductor Lp is connected to the virtual current source, the virtual current source is connected to the positive electrode of the floating charge power supply, and the negative electrode of the floating charge power supply is connected to the second end of the second inductor Ln; the multidimensional mapping model is connected to the virtual current source to control the output of the virtual current source.

Images