Method and system for evaluating reutilization of battery in battery changing cabinet

Abstract

The invention relates to the technical field of battery replacement cabinet batteries, and discloses a battery replacement cabinet battery recycling evaluation method and system, and the method comprises the steps: deploying a distributed detection device network in a battery replacement cabinet battery cluster, and collecting a battery original data set of the battery replacement cabinet battery cluster; performing feature classification and vector coding on the battery original data set to obtain a plurality of battery performance feature vectors; inputting the plurality of battery performance feature vectors into a battery reutilization potential prediction model set for prediction, and outputting a comprehensive battery reutilization potential prediction result; and setting evaluation node parameters based on the comprehensive battery reuse potential prediction result, and performing adaptive sampling and dynamic optimization calculation on the evaluation node parameters to obtain a battery multi-stage evaluation control scheme, so that comprehensive evaluation of the battery reuse potential is realized, the prediction accuracy is improved, and the generalization ability is relatively high.

Description

Technical Field

[0001] The present invention relates to the field of battery for power swap cabinets, and in particular to a method and system for evaluating the reuse of batteries for power swap cabinets. Background Art

[0002] During the service life of the battery in the battery swap cabinet, the battery performance will gradually deteriorate due to factors such as the increase in charge and discharge times and changes in ambient temperature. When key indicators such as capacity or power cannot meet the battery swap requirements, these batteries need to be retired. In order to maximize the use value of the battery, it is particularly important to conduct a reasonable reuse evaluation of retired batteries.

[0003] At present, conventional battery recycling assessment methods mainly rely on simple capacity tests and internal resistance tests, lacking systematic analysis and in-depth mining of battery life cycle data. This assessment method is difficult to accurately reflect the true health status of the battery, and it is also impossible to accurately assess the residual value of the battery. At the same time, due to the large number of battery packs in the battery swap cabinet, the traditional centralized assessment solution faces problems such as low data collection efficiency, heavy computing burden, and insufficient assessment accuracy in actual applications. Summary of the invention

[0004] The present invention provides a method and system for evaluating the reuse of batteries in a battery swap cabinet. The present invention realizes a comprehensive evaluation of the reuse potential of batteries, improves the prediction accuracy, and has strong generalization ability.

[0005] In a first aspect, the present invention provides a battery reuse assessment method for a battery swap cabinet, the battery reuse assessment method for a battery swap cabinet comprising:

[0006] Deploy a distributed detection device network in the battery cluster of the battery swap cabinet, and collect the battery original data set of the battery cluster of the battery swap cabinet;

[0007] Performing feature classification and vector encoding on the battery original data set to obtain multiple battery performance feature vectors;

[0008] Inputting the plurality of battery performance feature vectors into a battery reuse potential prediction model set for prediction, and outputting a comprehensive battery reuse potential prediction result;

[0009] The evaluation node parameters are set based on the comprehensive battery reuse potential prediction result, and the evaluation node parameters are adaptively sampled and dynamically optimized to obtain a multi-stage battery evaluation control scheme.

[0010] In a second aspect, the present invention provides a battery reuse evaluation system for a power exchange cabinet, the battery reuse evaluation system for a power exchange cabinet comprising:

[0011] A collection module, used to deploy a distributed detection device network in a battery cluster of a battery swap cabinet and collect the battery raw data set of the battery cluster of the battery swap cabinet;

[0012] An encoding module, used for performing feature classification and vector encoding on the battery original data set to obtain a plurality of battery performance feature vectors;

[0013] A prediction module, used for inputting the plurality of battery performance feature vectors into a battery reuse potential prediction model set for prediction, and outputting a comprehensive battery reuse potential prediction result;

[0014] The calculation module is used to set the evaluation node parameters based on the comprehensive battery reuse potential prediction result, and to perform adaptive sampling and dynamic optimization calculation on the evaluation node parameters to obtain a battery multi-stage evaluation control scheme.

[0015] The third aspect of the present invention provides a battery reuse evaluation device for a battery swap cabinet, comprising: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the battery reuse evaluation device for the battery swap cabinet executes the above-mentioned battery reuse evaluation method for the battery swap cabinet.

[0016] A fourth aspect of the present invention provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, it enables the computer to execute the above-mentioned battery reuse evaluation method for battery swap cabinets.

[0017] In the technical solution provided by the present invention, by deploying a distributed detection device network and combining a multi-level data collection and processing mechanism, the present invention realizes the efficient acquisition and analysis of the battery life cycle data of the battery swap cabinet, and significantly improves the integrity and accuracy of data collection. An evaluation method combining deep learning with variational Bayesian inference is adopted, and the accuracy of battery performance evaluation is improved through multi-level feature extraction and dynamic optimization strategies. The cascade utilization decision-making mechanism based on multi-objective optimization is combined with scene feature matching and resource optimization allocation strategies to improve the matching accuracy of battery reuse. Through adaptive sampling and dynamic optimization control, the present invention can automatically adjust the evaluation strategy according to changes in battery performance, introduce a multi-layer variational Bayesian inference network, and combine the analysis of battery performance degradation laws to improve the reliability of the evaluation results. By integrating multiple prediction sub-models and a dynamic weight allocation mechanism, the present invention realizes a comprehensive evaluation of the potential for battery reuse, improves the prediction accuracy, and has a strong generalization ability.

[0018] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description, claims and drawings.

[0019] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A schematic diagram of an embodiment of a battery reuse evaluation method for a battery swap cabinet in an embodiment of the present invention;

[0021] Figure 2 A schematic diagram of an embodiment of a battery reuse evaluation system for a battery swap cabinet in an embodiment of the present invention;

[0022] Figure 3 It is a schematic diagram of an embodiment of a battery reuse evaluation device for a battery swap cabinet in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] The terms "including" and "having" and any variations thereof mentioned in the embodiments of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device end including a series of steps or units is not limited to the listed steps or units, but may optionally include other steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or device ends.

[0025] To facilitate understanding of this embodiment, a battery recycling evaluation method for a battery swap cabinet disclosed in an embodiment of the present invention is first described in detail. Figure 1 As shown, the method comprises the following steps:

[0026] 101. Deploy a distributed detection device network in the battery cluster of the battery swap cabinet, and collect the original battery data set of the battery cluster of the battery swap cabinet;

[0027] It is understandable that the execution subject of the present invention may be a battery recycling evaluation system for a battery swap cabinet, or may be a terminal or a server, which is not specifically limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.

[0028] Specifically, a distributed detection device network is deployed in the battery cluster of the battery swap cabinet, and the battery cluster of the battery swap cabinet is grouped according to the geographical location information and the use environment parameters, and the distributed nodes are optimized based on these grouping information to form a distributed detection device network that can efficiently cover all batteries. In the process of building this network, the detection range and priority distribution of each node are determined by comprehensively analyzing the distribution characteristics of the geographical location, environmental conditions (such as temperature and humidity, altitude, etc.), battery usage status (such as charging and discharging frequency), and data collection requirements, so as to optimize the node layout and achieve efficient operation of the distributed network. In the deployed distributed detection device network, the operating parameters of the network nodes are synchronously calibrated, and unified synchronous sampling control parameters are set through the coordination of data sampling cycles and timestamp alignment technology to ensure that all distributed nodes collect data under the same time reference. Based on the synchronous sampling control parameters, the key physical parameters of the battery, including voltage, current, and temperature, are sampled at a high frequency to record the operating conditions of the battery. The collected battery operating condition data is input into the feature analysis module, and time series characteristics such as voltage fluctuation amplitude and temperature change rate are extracted through time domain feature analysis. At the same time, the frequency domain feature analysis is used to explore the hidden battery performance change rules, such as frequency response characteristics and harmonic distribution. The results of time domain and frequency domain feature analysis are combined to obtain a multi-dimensional description of the battery operating state and form the dynamic characteristic data of the battery. Based on the battery dynamic characteristic data, a voltage-current response curve is constructed, which intuitively reflects the intrinsic electrochemical performance of the battery by showing the relationship between the voltage and current changes of the battery under different working conditions. By analyzing the voltage-current response curve, the internal resistance change value of the battery is extracted using the relevant calculation model. While calculating the internal resistance change value, the temperature compensation coefficient is introduced to correct the result to improve the accuracy of the internal resistance parameter. Combined with the corrected internal resistance parameters, the performance degradation of the battery is evaluated. At the same time, the charging and discharging process of the batteries in the battery cluster of the power exchange cabinet at different use stages is recorded, and the charge and discharge capacity of the battery is accurately calculated by applying the charge integration method, and the number of battery cycles is counted by the coulomb counting method. Combining the dual calculation method of charge and discharge capacity and number of cycles, the battery life and remaining performance can be more accurately evaluated to form comprehensive battery usage record data. In order to achieve an overall evaluation of the battery operating status, the battery operating condition data, battery dynamic characteristic data, battery internal resistance parameters and battery usage record data are merged in time series. By aligning these data to a unified time axis, a consistent time series data set is formed. At the same time, in order to eliminate dimensional differences and noise in the data, data standardization processing technology is used to normalize all data to enhance the comparability and consistency between data, and finally obtain a battery original data set that can fully reflect the battery operating status.

[0029] 102. Perform feature classification and vector encoding on the original battery data set to obtain multiple battery performance feature vectors;

[0030] Specifically, based on the voltage parameters, current parameters and temperature parameters in the original data set of the battery, data dimension reduction processing is performed. By adopting dimension reduction techniques such as principal component analysis or linear discriminant analysis, high-dimensional multivariate data is converted into a low-dimensional representation to obtain a battery parameter dimension reduction matrix. The battery parameter dimension reduction matrix is ​​subjected to time series correlation analysis, the time series correlation between different parameters is quantified, and the time series feature weight coefficient is calculated to reflect the relative importance of each parameter to the battery performance at different time points. The time series feature weight coefficient is weighted and fused with the battery parameter dimension reduction matrix, and multiple first battery performance characteristics are extracted by combining the dynamic change relationship between the parameters. At the same time, in order to comprehensively evaluate the performance decay characteristics of the battery, the data related to the capacity in the original data set of the battery is decomposed into a time series, and the capacity decay curve is fitted by a double exponential function, so as to construct a capacity decay model that can describe the law of capacity change. In order to improve the accuracy of the fitting, the capacity decay curve is corrected in combination with the temperature stress coefficient so that the model can adapt to different usage environments. Through this process, the capacity decay model parameters reflecting the law of battery capacity change are obtained. Based on the capacity decay model parameters, a battery capacity prediction function is constructed. By using Taylor expansion to approximately expand the capacity prediction function, the capacity decay rate under different working conditions is calculated to obtain the capacity decay prediction value. At the same time, in order to analyze the internal resistance change characteristics of the battery, the internal resistance data in the original data is processed in segments, and the internal resistance change function is established based on the segmented linear regression method. On this basis, the parameters are optimized in combination with historical data to obtain the internal resistance change model coefficient that can reflect the law of internal resistance change. The internal resistance change model coefficient is input into the prediction function, and the internal resistance change value at different time points is predicted by recursive calculation. Due to the significant influence of temperature on the change of battery internal resistance, the temperature correction coefficient is used to correct the internal resistance prediction value to improve the accuracy and reliability of the prediction, and the corrected internal resistance change prediction value is obtained. The capacity decay prediction value and the internal resistance change prediction value are respectively subjected to feature vectorization processing, and these two types of characteristics are converted into an operable multidimensional feature space to generate multiple second battery performance characteristics. The first battery performance characteristics are combined with the second battery performance characteristics, and all extracted features are classified by feature classification algorithms (such as K-means clustering or support vector classification), and the classification results are vector encoded to generate multiple battery performance feature vectors.

[0031] A plurality of first battery performance characteristics and a plurality of second battery performance characteristics are normalized to eliminate the dimensional differences between different characteristics and the deviation of data distribution. The normalization process uses methods such as maximum and minimum value normalization or Z score normalization to map all feature values ​​into a unified numerical range, thereby improving the stability of feature calculation and cluster analysis. After the normalization process is completed, the feature space is analyzed using a density peak detection algorithm. By calculating the density distribution of data points and their distances to high-density areas, the initial value of the cluster center of the feature space is identified to obtain the first cluster centroid. The first cluster centroid is used as input and the K-means algorithm is applied for iterative optimization. The K-means algorithm takes minimizing the intra-class distance and maximizing the inter-class distance as the objective function. By repeatedly allocating data points and adjusting the centroid position, the clustering results are continuously optimized to obtain the second cluster centroid. Based on the second cluster centroid, a silhouette coefficient analysis is performed to quantify the quality of the current clustering result and determine the optimal number of cluster categories. The silhouette coefficient provides an objective basis for selecting a reasonable number of cluster categories by evaluating the distance relationship between data points within their own cluster and other clusters. Based on the optimal number of clustering categories, the battery performance characteristics are finally clustered. On the basis of clustering, the entropy method is used to calculate the information entropy and weight coefficient of each feature, evaluate the contribution of each feature to the overall feature space, and obtain the feature importance score. The entropy method ensures the scientificity and objectivity of weight allocation by measuring the distribution differences of features between samples. The feature importance score is weighted and combined with the clustering results to generate the final representation of the feature. In order to optimize the feature encoding strategy, the local sensitive hash function is used to generate the hash coding strategy. The local sensitive hash function maintains the relative similarity between features by efficiently mapping high-dimensional features to low-dimensional space, while reducing the computational complexity. The features are mapped to a reduced dimension by the hash coding strategy to obtain the correlation strength of the features. Based on the feature correlation strength, high-order features are extracted and combined for multiple first battery performance features and multiple second battery performance features. Through nonlinear feature combination and interaction effect analysis, the potential high-order relationships in the original feature space are mined to construct a richer feature expression form. On the basis of high-order feature extraction, multiple battery performance feature vectors are finally formed through reasonable feature fusion technology.

[0032] 103. Input multiple battery performance feature vectors into a battery reuse potential prediction model set for prediction, and output a comprehensive battery reuse potential prediction result;

[0033] Specifically, by performing subspace partitioning on multiple battery performance feature vectors, the high-dimensional feature space is decomposed into several independent feature subspaces. Feature selection and dimensionality reduction techniques, such as principal component analysis, linear discriminant analysis or t-SNE, are used to ensure that the divided feature subspaces are independent of each other while retaining sufficient feature information. Through subspace partitioning, complex high-dimensional features are effectively decomposed into multiple low-dimensional subspaces, thereby improving the computational efficiency and prediction ability of the model. The independent feature subspaces are respectively input into the battery reuse potential prediction model set for prediction. The model set consists of multiple battery reuse potential prediction submodels, each of which has an independent structure and is optimized for a specific subspace. The prediction submodel consists of three parts: a feature extraction layer, a feature mapping layer, and a prediction output layer. Among them, the feature extraction layer consists of three fully connected layers, each of which uses BatchNormalization and ReLU activation functions to ensure balanced distribution of feature data and enhanced nonlinear modeling capabilities. BatchNormalization technology can stabilize the training process and accelerate model convergence, while the ReLU activation function improves the model's ability to characterize complex feature relationships by introducing nonlinearity. The feature mapping layer in the feature extraction layer consists of two residual connection blocks, each of which contains a two-layer perceptron and a jump connection. The residual connection design alleviates the gradient vanishing problem caused by the deep model through a direct short-circuit path, while enhancing the network's sensitivity to input features and learning ability. The prediction output layer uses a Sigmoid activation function to map the prediction results of the sub-model to between 0 and 1, which facilitates the probabilistic interpretation of the battery reuse potential. Each sub-model outputs a prediction result based on the corresponding independent feature subspace, and the prediction results of all sub-models together constitute a sub-model prediction result set. Dynamic weights are assigned to the sub-model prediction results to optimize the final comprehensive prediction results. The weight assignment process is based on a dynamic weighting method, combining the historical performance and current prediction accuracy of each sub-model, and dynamically calculating the fusion weight coefficient through Bayesian optimization or a weight adjustment strategy based on the attention mechanism. Based on the fusion weight coefficient, a weighted analysis is performed on the sub-model prediction result set, and the prediction results of each sub-model and its corresponding weight are weighted summed to generate a comprehensive battery reuse potential prediction result. 104. The evaluation node parameters are set based on the comprehensive battery reuse potential prediction results, and the evaluation node parameters are adaptively sampled and dynamically optimized to obtain a multi-stage battery evaluation control scheme.

[0034] Specifically, the comprehensive battery reuse potential prediction results are input into the first-layer variational Bayesian inference network for data distribution modeling. Through this process, the variational expectation maximization algorithm is used to iteratively calculate the hidden variable posterior distribution of the data to derive the distribution parameters of the battery performance. The variational Bayesian inference network handles complex and high-dimensional battery performance data distribution problems through its nonlinear modeling capabilities, so that the calculated distribution parameters can accurately reflect the actual situation. The battery performance distribution parameters are input into the second-layer variational Bayesian inference network to analyze the evolution of battery performance. Combined with random field theory, the network calculates the transition probability matrix of the battery state to describe the probability distribution of the transition of battery performance from one state to another. Through the state transition probability matrix, the inherent law of battery performance degradation is revealed, providing a basis for evaluating its performance evolution trend throughout its life cycle. Probabilistic modeling of state transition through random field theory can adapt to the nonlinear and random characteristics of battery performance changes and ensure the reliability and scientificity of the analysis results. The battery performance degradation law is input into the third-layer variational Bayesian inference network to perform probabilistic modeling of the evaluation node parameters. By refining the distribution of node parameters, the network outputs a set of high-precision evaluation node parameters. These parameters reflect the key performance indicators of the battery at different stages, including capacity, internal resistance, remaining life, etc. Based on the evaluation node parameters, a multi-stage evaluation control strategy is designed and optimized. The optimization process comprehensively considers the dynamics of battery performance characteristics, and generates an evaluation optimization control strategy by establishing a mapping relationship between the evaluation node and the battery state. In order to improve the applicability and flexibility of the evaluation strategy, the evaluation optimization control strategy is adaptively sampled to obtain an adaptive sampling scheme. By dynamically adjusting the sampling frequency and sampling range, the evaluation strategy can adapt to changes in battery performance in real time. The adaptive sampling scheme is subjected to time series feature extraction to obtain a dynamic adjustment matrix of evaluation parameters. The dynamic adjustment matrix quantifies the trend of battery performance over time and provides a basis for parameter adjustment for multi-stage evaluation. Based on this matrix, the evaluation accuracy and evaluation efficiency are jointly optimized to ensure the scientificity and practicality of the evaluation scheme. In the joint optimization process, a multi-objective optimization method is adopted, and the improvement of evaluation accuracy and the reduction of resource consumption are taken as optimization goals, and the balance between the two is achieved through iterative adjustment. Combining the outputs of all steps, a complete multi-stage battery evaluation control scheme is generated. The solution integrates historical data and future trends of battery performance, and can dynamically adjust evaluation strategies to adapt to changes in battery performance, thereby achieving scientific management of the entire life cycle of the battery.

[0035] In the embodiment of the present invention, by deploying a distributed detection device network and combining a multi-level data collection and processing mechanism, the present invention realizes the efficient acquisition and analysis of the battery life cycle data of the battery swap cabinet, and significantly improves the integrity and accuracy of data collection. An evaluation method combining deep learning with variational Bayesian inference is adopted, and the accuracy of battery performance evaluation is improved through multi-level feature extraction and dynamic optimization strategies. The cascade utilization decision-making mechanism based on multi-objective optimization is combined with scene feature matching and resource optimization allocation strategy to improve the matching accuracy of battery reuse. Through adaptive sampling and dynamic optimization control, the present invention can automatically adjust the evaluation strategy according to changes in battery performance, introduce a multi-layer variational Bayesian inference network, and combine the analysis of battery performance degradation laws to improve the reliability of the evaluation results. By integrating multiple prediction sub-models and a dynamic weight allocation mechanism, the present invention realizes a comprehensive evaluation of the potential for battery reuse, improves the prediction accuracy, and has a strong generalization ability.

[0036] In a specific embodiment, the process of executing step 101 may specifically include the following steps:

[0037] The battery cluster of the battery swap cabinet is grouped and the distributed nodes are optimized according to the geographical location information and the use environment parameters to obtain a distributed detection device network;

[0038] Performing data sampling cycle synchronization and time stamp alignment on a distributed detection device network to obtain synchronous sampling control parameters;

[0039] Based on the synchronous sampling control parameters, the battery voltage parameters, current parameters and temperature parameters are sampled at high frequency to obtain the battery operating condition data, and the battery operating condition data is analyzed in the time domain and frequency domain to obtain the battery dynamic characteristic data;

[0040] Based on the dynamic characteristic data of the battery, a voltage-current response curve is constructed, and the change value of the battery internal resistance is calculated through the voltage-current response curve. At the same time, correction is performed in combination with the temperature compensation coefficient to obtain the battery internal resistance parameter;

[0041] Record the charging and discharging process of the batteries in the battery cluster of the power exchange cabinet, calculate the charging and discharging capacity by the charge integration method, and count the number of cycles by combining the Coulomb counting method to obtain the battery usage record data;

[0042] The battery operating condition data, battery dynamic characteristic data, battery internal resistance parameters and battery usage record data are time-series merged and data standardized to obtain the battery original data set.

[0043] Specifically, the battery cluster of the battery swap cabinet is grouped and the distributed nodes are optimized according to the geographic location information and the use environment parameters to establish an efficient distributed detection device network. Assume that the battery swap cabinet network of a city is distributed in multiple areas, and the geographic location information of each area is represented by coordinates (x i ,y i ), where i is the number of the power exchange cabinet. At the same time, environmental parameters such as temperature (T i )、Humidity (H i ) and load frequency (f i ). In order to optimize the distributed nodes, construct the objective function where w i represents the weight, d i is the distance between the battery swap cabinets, λ is the regularization parameter, P i Represents the load parameter of the node. By solving this optimization problem, the power exchange cabinets are effectively grouped and the optimal deployment location of the detection device is determined to ensure the maximization of network coverage and data collection efficiency. After building a distributed detection device network, the data sampling cycle synchronization and timestamp alignment are performed to ensure that the data from different detection nodes are consistent on the same time basis. Assume that the sampling period of each node is T s , timestamp is t i , synchronous sampling control parameters through Δt = t i -t j (where i and j are different nodes) are adjusted so that all nodes satisfy t i +kT s =t j +mT s , where k,m∈Z. In this way, the data offset between different nodes is eliminated and global synchronization is achieved. After synchronization, the key parameters of the battery in the battery swap cabinet are sampled at high frequency based on the control parameters, including voltage (V), current (I) and temperature (T). Let the sampling frequency be f s , then the sampled data at time t is expressed as {V(t), I(t), T(t)}. After the collected battery operating data is analyzed by time domain characteristics and frequency domain characteristics, key dynamic characteristics are extracted. For example, the time domain characteristics of voltage and current are calculated by averaging and standard deviation (where X represents V or I). In the frequency domain, fast Fourier transform is used to analyze the spectral characteristics of the signal and extract the main frequency component and high-frequency harmonic information. These analysis results are used to construct a battery dynamic characteristic data set. Based on the battery dynamic characteristic data, the voltage-current response curve is drawn. Assume that the sampling data of voltage and current are The response curve is obtained by fitting the formula V = RI + V0 (where R is the internal resistance and V0 is the open circuit voltage). By optimizing the least squares method, the Since temperature has a significant impact on internal resistance, a temperature compensation coefficient α is introduced. T Correct the internal resistance value and obtain the corrected internal resistance parameter R T =R(1+α T (TT ref )), where T ref is the reference temperature. For the record of the charge and discharge process, the charge and discharge capacity is calculated by the charge integration method. The charge capacity Q of the battery in time [t1, t2] c And the discharge capacity Q d They are and Combined with the Coulomb counting method, the number of battery cycles n is counted, and the calculation formula is: Where Q rated is the rated capacity, and k is the cycle number. The battery operating data, battery dynamic characteristic data, battery internal resistance parameters, and battery usage record data are merged in time series, and the data of different time axes are unified to the global time axis by interpolation method, and the merged data are standardized. For example, the minimum-maximum normalization is used to map the data to the range of [0,1], and the normalization formula is The standardized data set is the original battery data set, which can fully reflect the performance status of the battery at different times and working conditions.

[0044] In a specific embodiment, the process of executing step 102 may specifically include the following steps:

[0045] Based on the original battery data set, the voltage parameters, current parameters and temperature parameters are reduced in dimension to obtain the battery parameter reduction matrix, and the battery parameter reduction matrix is ​​subjected to time series correlation analysis to obtain the time series feature weight coefficient;

[0046] Performing a weighted fusion operation on the time series feature weight coefficient and the battery parameter dimension reduction matrix to obtain a plurality of first battery performance features;

[0047] The capacity data in the original battery data set is decomposed into time series, and the capacity decay curve is fitted by a double exponential function. The capacity decay model parameters are obtained by combining the temperature stress coefficient for correction.

[0048] The battery capacity prediction function is constructed based on the capacity decay model parameters, and the capacity decay rate under different working conditions is calculated by Taylor expansion to obtain the capacity decay prediction value;

[0049] The internal resistance data in the original battery data set is processed in segments, and the internal resistance change function is established by the segmented linear regression method. The parameters are optimized in combination with historical data to obtain the internal resistance change model coefficients.

[0050] The internal resistance change model coefficient is input into the prediction function, and the internal resistance prediction value at different time points is obtained through recursive calculation, and the internal resistance change prediction value is obtained by correction combined with the temperature correction coefficient;

[0051] Performing feature vector quantization processing on the capacity attenuation prediction value and the internal resistance change prediction value to obtain a plurality of second battery performance characteristics;

[0052] Feature classification and vector encoding are performed on the plurality of first battery performance features and the plurality of second battery performance features to obtain a plurality of battery performance feature vectors.

[0053] Specifically, the voltage parameter V(t), current parameter I(t) and temperature parameter T(t) in the original battery data set are reduced in dimension. Assuming that the data set contains N samples, X=[V,I,T] is an N×3 matrix, where each column represents the time series of a parameter. Through principal component analysis, the dimension reduction matrix Z=XW is calculated, where W is the eigenvector matrix, which defines the projection direction of the original parameters in the reduced dimension space. The reduced dimension matrix Z retains the main features of the voltage, current and temperature parameters, while reducing the data dimension to reduce redundancy. Perform time series correlation analysis on the battery parameter dimension reduction matrix Z and extract the time series feature weight coefficients. Let Z=[z1,z2,...,z m ] represents the feature component after dimensionality reduction, and the autocorrelation function is calculated for each component And determine the time correlation weight of the feature component according to different lag steps τ. Combine these weight coefficients to construct the time series feature weight vector w t The time series feature weight coefficient is weighted and fused with the dimension reduction matrix Z. The calculation formula is F1 = w t T Z, and obtain multiple first battery performance characteristics. The capacity data C(t) in the original battery data set is decomposed into a time series, and its change trend is modeled by fitting the capacity decay curve with a double exponential function. The capacity decay is expressed as C(t) = C0e -αt +C r e -βt , where C0 is the initial capacity, α and β are the decay rate constants, and C r is the residual capacity. The nonlinear least squares method is used to optimize the fitting parameters α, β, C0 and C r Since temperature has a significant effect on capacity attenuation, combined with the temperature stress coefficient γ T , the modified capacity model is C T (t) = C (t) (1 + γ T (TT ref )), where T ref Is the reference temperature. Based on the capacity decay model parameters, a battery capacity prediction function is constructed. The capacity decay rate under different working conditions is approximately calculated by Taylor expansion. The expression of the prediction function is C p (t)≈C T (t)-(αC0e -αt +

[0054] βC r e -βt )Δt. The prediction function is used to derive the predicted value of capacity decay in the future, and then evaluate the battery life and remaining performance. At the same time, the internal resistance data R(t) in the original battery data set is segmented to capture the changing trend of the internal resistance at different stages. Assume that the internal resistance data is divided into k stages, and the internal resistance trend of each stage is expressed as R(t)=a i t+b i ,t∈[t i-1 ,t i ], where a i and b i is the linear regression coefficient of stage i. i ,b i ) and optimize the model in combination with historical data to obtain the complete internal resistance change model coefficients. These coefficients are input into the internal resistance prediction function R p (t) = R (t) + ΔR T , where ΔR T =γ R (TT ref ) represents the temperature correction term, and the final predicted value of internal resistance change R is obtained p (t). The predicted value of capacity attenuation and the predicted value of internal resistance change are feature vectorized to generate multiple second battery performance characteristics. These characteristics are uniformly described in a high-dimensional representation of the dynamic changes in battery performance in the two dimensions of capacity and internal resistance. The first battery performance characteristic F1 and the second battery performance characteristic F2 are feature classified and vector encoded. All features are clustered and classified using K-means or density peak clustering algorithms to obtain feature classification labels. Combined with these labels, the features are vector encoded, for example, the features are embedded in a low-dimensional space to form the final battery performance feature vector.

[0055] In a specific embodiment, the process of performing feature classification and vector encoding on the plurality of first battery performance features and the plurality of second battery performance features to obtain the plurality of battery performance feature vectors may specifically include the following steps:

[0056] Normalizing the plurality of first battery performance characteristics and the plurality of second battery performance characteristics, calculating an initial value of a feature space cluster center by a density peak detection algorithm, and obtaining a first cluster centroid;

[0057] The first cluster centroid is input into the K-means algorithm, and the second cluster centroid is obtained by iterative calculation through the optimization objective function of minimizing the intra-class distance and maximizing the inter-class distance;

[0058] The silhouette coefficient analysis is performed on the second cluster centroid to obtain the optimal number of cluster categories, and cluster division is performed based on the optimal number of cluster categories. The information entropy and weight coefficient of each feature are calculated by the entropy method to obtain the feature importance score;

[0059] The feature importance score is weightedly combined with the clustering result, a hash coding strategy is generated through a local sensitive hash function, and dimensionality reduction mapping is performed according to the hash coding strategy to obtain the feature association strength;

[0060] Based on the feature correlation strength, high-order features are extracted and combined for the plurality of first battery performance features and the plurality of second battery performance features to obtain a plurality of battery performance feature vectors.

[0061] Specifically, for a plurality of first battery performance characteristics F1=[f 11 ,f 12 ,…,f 1m ] and a plurality of second battery performance characteristics F2 = [f 21 ,f 22 ,…,f 2n ] to perform normalization, using the maximum and minimum normalization formula Map all eigenvalues ​​to the range of [0,1] to eliminate the dimensional differences between eigenvalues ​​and improve the accuracy of subsequent clustering and analysis. Perform an initial estimate of the cluster center in the feature space based on the density peak detection algorithm. Assume that the normalized feature matrix is ​​F = [F1; F2], where each row represents a feature sample and each column represents a normalized feature. In the density peak detection algorithm, the distance matrix D = [d ij ], where d ij =||f i -f j || represents the Euclidean distance between samples i and j. Define the local density of each sample Where δ is the distance kernel width parameter. At the same time, the minimum distance δ from each sample to a higher density sample is calculated i =min j:ρj>ρi (d ij ). i and δ i Combined, select the sample with the highest comprehensive score of density and distance as the cluster center to obtain the first cluster centroid. Input the first cluster centroid as the initial condition into the K-means clustering algorithm to optimize the cluster center. K-means minimizes the sum of squares of intra-class distances. The objective function iteratively adjusts the cluster center c k , and redistribute samples to the nearest cluster center in each iteration until the objective function converges. After iterative optimization, the second cluster centroid is obtained, which represents a more accurate clustering result. The silhouette coefficient analysis is performed on the second cluster centroid to determine the optimal number of cluster categories K * The silhouette coefficient S(i) is defined as Where a(i) is the average distance between sample i and other samples in its cluster, and b(i) is the average distance between sample i and the nearest cluster. By calculating the average silhouette coefficient of all samples, K is selected. * Make the silhouette coefficient reach the maximum value, and based on K * After the clustering is completed, the entropy method is used to calculate the information entropy and weight coefficient of each feature. The information entropy formula is H(f j )=-∑ i p ij log(p ij ),in Represents the normalized probability distribution of feature j. The weight coefficient is calculated based on information entropy. Get the feature importance score. These scores are weighted and combined with the clustering results to generate the final representation of the feature. In order to optimize feature processing and reduce dimensionality, a local sensitive hash function is used to generate a hash coding strategy. The feature is mapped to a reduced dimension using the local sensitive hash function. The mapping formula is: where a k is a random vector, b k is the offset, and the generated hash code achieves efficient dimensionality reduction by retaining the local similarity of similar feature samples. The feature association strength matrix S after dimensionality reduction represents the similarity relationship between features. Based on the feature association strength, high-order features are extracted and combined for the first battery performance feature and the second battery performance feature. High-order features are extracted through interaction terms and multi-layer nonlinear transformations to generate multiple battery performance feature vectors. For example, assuming that the feature vector is f i =[f i1 ,f i2 ,f i3 ], by generating the quadratic interaction term f ij ·f ik and nonlinear mapping φ(f i j)=ReLU(Wf ij +b), forming a high-order feature matrix.

[0062] In a specific embodiment, the process of executing step 103 may specifically include the following steps:

[0063] Perform subspace division based on multiple battery performance feature vectors to obtain multiple independent feature subspaces;

[0064] Input multiple independent feature subspaces into a battery reuse potential prediction model set, the battery reuse potential prediction model set includes multiple battery reuse potential prediction sub-models, the battery reuse potential prediction sub-models include a feature extraction layer, a feature mapping layer and a prediction output layer, the feature extraction layer includes 3 fully connected layers, each fully connected layer uses BatchNormalization and ReLU activation function, the feature mapping layer includes 2 residual connection blocks, each residual connection block includes a two-layer perceptron and a skip connection, and the prediction output layer uses a Sigmoid activation function;

[0065] The independent feature subspaces are respectively input into the corresponding battery reuse potential prediction sub-models to obtain a sub-model prediction result set, and the sub-model prediction result set is dynamically weighted to obtain a fusion weight coefficient;

[0066] The sub-model prediction result set is weighted analyzed based on the fusion weight coefficient to output the comprehensive battery reuse potential prediction result.

[0067] Specifically, based on multiple battery performance feature vectors F = [f1, f2, ..., f N ], the high-dimensional feature space is decomposed into multiple independent feature subspaces {S1, S2, …, S K The subspace partitioning is done by principal component analysis or kernel method. For example, principal component analysis calculates the covariance matrix The eigenvalues ​​and eigenvectors of , project the original features to the first K principal components, and form a low-dimensional subspace matrix S k =XW k , where W k is the projection matrix of the corresponding subspace. The independent feature subspaces are input into the battery reuse potential prediction model set. The model set consists of multiple submodels, each of which focuses on an independent feature subspace. Each prediction submodel consists of a feature extraction layer, a feature mapping layer, and a prediction output layer. In the feature extraction layer, three fully connected layers are used, and the calculation formula for each layer is z (l) =ReLU(BatchNorm(W (l) z (l-1) +b (l ))), where z (l) is the output of layer l, W (l) and b (l)are weight matrix and bias vector respectively, BatchNorm represents batch normalization operation, and ReLU(x)=max(0,x) is activation function. This structure can effectively extract nonlinear relationship of subspace features and improve training stability through normalization. The feature mapping layer consists of 2 residual connection blocks, each of which includes a two-layer perceptron and a skip connection. Assume that the input is h in , the output of the residual block is h out =h in +MLP(h in ), where MLP(h)=ReLU(W2ReLU(W1h+b1)+b2) is a two-layer perceptron. This structure alleviates the gradient vanishing problem through skip connections and enhances feature extraction capabilities. In the prediction output layer, the Sigmoid activation function is used to map the output value of the model to the [0,1] interval and output the reuse potential prediction value of each sub-model in is the prediction result of the kth sub-model. By inputting all independent feature subspaces into the corresponding prediction sub-models, we can get the sub-model prediction result set. Dynamically weight the set of sub-model prediction results to reflect the relative importance of different sub-models. The weight allocation is done by the weighted formula Calculate, where MSE k is the mean square error of the kth sub-model on the validation data, ω k is the corresponding fusion weight coefficient. The dynamic weight allocation method can automatically adjust the weight so that the sub-model with better performance occupies a larger proportion in the comprehensive prediction. Based on the fusion weight coefficient, the sub-model prediction result set is weighted and analyzed to calculate the comprehensive battery reuse potential prediction result. The calculation formula for the comprehensive prediction result is in Represents the final comprehensive prediction result. The weighted method achieves a comprehensive evaluation of the battery reuse potential by integrating the prediction values ​​of multiple sub-models.

[0068] In a specific embodiment, the process of executing step 104 may specifically include the following steps:

[0069] The comprehensive battery reuse potential prediction results are input into the first-layer variational Bayesian inference network for data distribution modeling, and the posterior distribution of latent variables is calculated through the variational expectation maximization algorithm to obtain the battery performance distribution parameters;

[0070] The battery performance distribution parameters are input into the second-layer variational Bayesian inference network to analyze the battery performance evolution. The state transition probability matrix is ​​calculated by combining the random field theory to obtain the battery performance degradation law.

[0071] The battery performance degradation law is input into the third-layer variational Bayesian inference network, and the evaluation node parameters are probabilistically modeled to obtain the evaluation node parameters;

[0072] Based on the evaluation node parameters, multi-stage evaluation optimization is performed to obtain the evaluation optimization control strategy, and the evaluation optimization control strategy is adaptively sampled to obtain an adaptive sampling scheme;

[0073] The timing features of the adaptive sampling scheme are extracted to obtain the dynamic adjustment matrix of evaluation parameters. Based on the dynamic adjustment matrix of evaluation parameters, the evaluation accuracy and evaluation efficiency are jointly optimized to obtain a multi-stage evaluation control scheme for the battery.

[0074] Specifically, the comprehensive battery reuse potential prediction result y is input into the first layer variational Bayesian inference network for data distribution modeling. Assume that the generation of y is controlled by the latent variable z, and its joint probability distribution is p(y,z)=p(y|z)p(z). In order to estimate the posterior distribution q(z) of the latent variable, the variational Bayesian framework is adopted to approximate the posterior distribution as q(z)=Π i q i (z i ), and iteratively optimize the variational parameters through the variational expectation maximization (EM) algorithm. E steps to calculate the variational lower bound L(q) = E q(z) [logp(y,z)]-E q(z) [logq(z)], the Mth step updates the model parameters by maximizing L(q) to obtain the battery performance distribution parameters θ = {μ, Σ}, where μ and Σ are the mean and covariance of the latent variables, respectively. The battery performance distribution parameters are input into the second layer variational Bayesian inference network to analyze the evolution of battery performance. Combined with random field theory, it is assumed that the battery performance state is represented as a time series {s t}, where the transition of each state is represented by the conditional probability p(s t+1 |s t ) control. Using the battery performance distribution parameter θ, through random field modeling, the state transition probability matrix P is defined, and its element P ij =p(s t+1 =j|s t =i) represents the transition probability from state i to state j. The expected value E[P ij], and obtain the law of battery performance degradation. This law describes the evolution trend of battery performance in the form of state transition probability. In the third-layer variational Bayesian inference network, the battery performance degradation law is input into the model, and the evaluation node parameters are probabilistically modeled. The posterior distribution of the evaluation node parameter q(p) is optimized by combining the performance degradation law and the node observation data. Assuming that the prior distribution of the node parameter is p(p), and the observation data is o, its posterior distribution is q(p)∝p(o|p)p(p). After optimization by variational inference, the probability representation of the node parameter p={p1,p2,…,p k}, each node parameter represents the evaluation weight of battery performance at different stages. Based on the evaluation node parameters, the multi-stage battery evaluation is optimized to generate an evaluation optimization control strategy. The optimization objective function is defined as where s t is the state distribution at stage t. The purpose of optimization is to maximize the overall evaluation benefit. The optimal control strategy is solved by gradient descent method or dynamic programming algorithm. Adaptive sampling is performed on the evaluation optimization control strategy. By adjusting the sampling frequency and strategy, an adaptive sampling scheme is generated. Assuming the sampling strategy is π(t), the optimization problem is defined as minimizing the sampling cost. where c t is the sampling cost of stage t, while satisfying the evaluation accuracy constraint Through the dynamic optimization method, an adaptive sampling scheme that can balance the sampling cost and evaluation accuracy is generated. After the adaptive sampling scheme is generated, the timing features are extracted to generate the evaluation parameter dynamic adjustment matrix A. The elements of the matrix A are ij It represents the dynamic adjustment weight of parameter j at time i. The specific calculation formula is: The weights of the evaluation parameters are dynamically adjusted through this matrix. Based on the dynamic adjustment matrix A of the evaluation parameters, the evaluation accuracy and evaluation efficiency are jointly optimized to obtain the battery multi-stage evaluation control scheme. The optimization objective is defined as O = α·Precision(p)-β·Cost(p), where α and β are trade-off coefficients, and the accuracy and cost are respectively expressed as Precision(p) = E[||p true -p|| 2 ],Cost(p)=∑ t c t The optimal evaluation parameter configuration is solved through multi-objective optimization algorithms, such as genetic algorithms or particle swarm optimization.

[0075] Among them, the battery performance degradation law is input into the third-layer variational Bayesian inference network, and the evaluation node parameters are probability modeled to obtain the evaluation node parameters; multi-stage evaluation optimization is performed based on the evaluation node parameters to obtain the evaluation optimization control strategy, and the evaluation optimization control strategy is adaptively sampled to obtain the adaptive sampling scheme, including: segmenting the battery performance degradation law, analyzing the battery capacity attenuation rate through the first battery performance evaluation unit to obtain the first evaluation parameter, analyzing the internal resistance change rate through the second battery performance evaluation unit to obtain the second evaluation parameter, and obtaining battery performance segmentation data; based on the battery performance segmentation data, the first evaluation parameter and the second evaluation parameter are input into the parameter association processing unit, and the mutual influence between the parameters is quantitatively analyzed to obtain the parameter correlation index; constructing an evaluation reference model for the parameter correlation index, obtaining the first optimization target value through quantitative indicator analysis in the first evaluation interval, and obtaining the second optimization target value through qualitative indicator analysis in the second evaluation interval, to the evaluation reference standard; input the evaluation reference standard into the multi-stage evaluation control unit, and obtain the evaluation control boundary conditions by setting the battery performance attenuation threshold, the evaluation time threshold and the resource consumption threshold; prioritize the first optimization target value and the second optimization target value based on the evaluation control boundary conditions, obtain the first control parameter through multi-objective collaborative optimization, obtain the second control parameter through hierarchical progressive optimization, and obtain the optimization control strategy; input the optimization control strategy into the adaptive sampling unit, dynamically adjust the first control parameter in the time dimension to obtain the first sampling sequence, and dynamically adjust the second control parameter in the space dimension to obtain the second sampling sequence; establish an evaluation quality feedback mechanism based on the first sampling sequence and the second sampling sequence, calculate the sampling accuracy through the first evaluation unit to obtain the quality score, and calculate the sampling efficiency through the second evaluation unit to obtain the resource occupancy rate; input the quality score and the resource occupancy rate into the optimization adjustment unit, perform online correction and dynamic compensation on the sampling parameters, and obtain the adaptive sampling scheme through feedback iterative operation.

[0076] In a specific embodiment, the method for evaluating battery reuse in a battery swap cabinet further includes the following steps:

[0077] The target layer of the multi-stage battery evaluation control scheme is constructed, and the three primary indicators of capacity decay rate, internal resistance change rate and cycle life are set through the hierarchical analysis method to obtain the evaluation index system;

[0078] According to the evaluation index system, a judgment matrix is ​​constructed, the importance of the indicators is compared pairwise, and the matrix is ​​normalized to obtain the indicator weight vector;

[0079] The consistency test of the indicator weight vector is carried out, and the consistency ratio is solved by the eigenvalue calculation method. When the consistency ratio is less than the preset target value, the target weight coefficient is obtained;

[0080] Perform weighted calculation on battery performance based on target weight coefficients to obtain battery grading results, and perform scene feature matching based on the battery grading results to obtain a scene matching matrix;

[0081] The optimal solution is searched through the scenario matching matrix to obtain the scenario priority ranking. Based on the scenario priority ranking, the constraint condition check is performed to ensure that the allocation plan meets the performance requirements and safety standards of each scenario, and the battery reuse execution plan is obtained.

[0082] Specifically, the target layer of the multi-stage battery evaluation control scheme is constructed, and three primary indicators are set: capacity decay rate, internal resistance change rate, and cycle life. These three indicators represent the rate of battery performance degradation, the impact of internal resistance changes on battery efficiency and heat generation, and the actual service life of the battery, and can comprehensively characterize the health status of the battery. In the hierarchical analysis method, when constructing an evaluation index system, it is necessary to clarify the target layer, criterion layer, and solution layer. The target layer is the battery reuse potential assessment, the criterion layer is the three primary indicators, and the solution layer is the actual optional scenario. According to the set evaluation index system, a judgment matrix A=[a ij ], where a ij Indicates the importance of indicator i relative to indicator j. After construction, the judgment matrix is ​​a positive reciprocal matrix that satisfies a ij =1 / a ji And a ii =1. For the three first-level indicators, the judgment matrix is ​​as follows:

[0083]

[0084] Through matrix normalization operation, calculate the weight vector w = [w1, w2, w3] of each indicator T The normalization process involves summing the elements of each column and then dividing each element by the sum of the column to obtain the normalized matrix A′=[a′ ij ],in After normalization, the mean of each row is calculated to obtain the weight vector Where n is the number of indicators. Perform consistency check on the calculated indicator weight vector to ensure the rationality of the judgment matrix. Calculate the consistency ratio and solve the maximum eigenvalue λ of the judgment matrix. max , through the formula Calculate, where (Aw) i is the i-th element of the matrix-vector product. Then the consistency index (Cl) is calculated as The eventual consistency ratio is RI is the random consistency index, and its value depends on the order of the matrix. When CR < 0.1, the consistency of the judgment matrix is ​​considered acceptable, and the weight vector is the target weight coefficient. Based on the target weight coefficient, the battery performance is weighted and the battery classification result is obtained. Assume that the battery performance index matrix is ​​P = [p ij ], where p ij It represents the value of the ith battery on the jth index. The battery classification result S = [s1, s2, ..., s m ]Through the formula i =∑ j w j p ij Calculate, where s i represents the comprehensive performance score of the ith battery. Based on the battery grading results, scene feature matching is performed to generate a scene matching matrix. Assume that the scene feature matrix is ​​C = [c kj ], where c kj represents the demand weight of scenario k for indicator j, then the scenario matching matrix M = [m ik ]Through the formula m ik =∑ j p ij c kj Calculate, where m ik Indicates the matching degree between battery i and scenario k. Using the scenario matching matrix, the optimization algorithm is used to search for the optimal solution to determine the scenario priority. Define the optimization objective function Q = ∑ i,k x ik m ik , where x ik is an allocation variable. If battery i is allocated to scenario k, then x ik =1, otherwise x ik = 0. The optimization process needs to meet constraints, for example, each battery can only be assigned to one scenario: And the sum of battery performance in each scenario is not less than the safety requirements: where τ k is the performance threshold of scenario k. Solve the optimization problem through linear programming or integer programming to obtain the scenario priority ranking. According to the scenario priority ranking and constraint condition verification, ensure that the generated battery reuse execution plan meets all performance requirements and safety standards, and obtain the battery reuse execution plan.

[0085] After obtaining the battery recycling execution plan, it also includes:

[0086] The batteries in the battery reuse implementation scheme are graded for cascade utilization, and the batteries are divided into a Class A cascade utilization group, a Class B cascade utilization group, and a recycling treatment group by analyzing the remaining capacity, internal resistance growth rate, and number of cycles of the batteries, to obtain the cascade grading results; based on the cascade grading results, a backup power supply adaptability test is performed on the batteries in the Class A cascade utilization group, and the adaptation parameters of the 48V communication base station backup power supply are obtained through 1C rate constant current discharge and 0.5C rate constant current charging tests; a low-voltage energy storage system adaptability test is performed on the batteries in the Class A cascade utilization group, and the adaptation parameters of the 220V household energy storage system are obtained through 0.3C rate constant power charge and discharge and voltage consistency tests; a photovoltaic energy storage adaptability test is performed on the batteries in the Class B cascade utilization group, and the 0.2C rate intermittent charge and discharge and environmental adaptability tests are obtained. Adaptation parameters of photovoltaic energy storage system; pre-treatment of the recycling group batteries before disassembly, and obtaining disassembly safety parameters through forced discharge, insulation test and pressure relief treatment; automatic disassembly of the recycling group batteries, and obtaining four types of recovery components, namely positive electrode materials, negative electrode materials, diaphragms and electrolytes through mechanical separation process and material sorting; extraction of metal elements from positive electrode materials, separation and purification of nickel, cobalt and manganese through wet smelting process to obtain battery metal raw materials, and graphite purification of negative electrode materials to obtain carbon material raw materials; chemical treatment of diaphragms and electrolytes, and obtaining renewable organic materials through solvent extraction process and high-temperature cracking process; establishment of recycling and utilization archives for battery metal raw materials, carbon material raw materials and renewable organic materials, and determination of recovery rate indicators through material balance calculation to form a closed-loop recycling system.

[0087] The above describes the battery reuse evaluation method of the battery swap cabinet in the embodiment of the present invention. The following describes the battery reuse evaluation system of the battery swap cabinet in the embodiment of the present invention. Figure 2 , an embodiment of the battery reuse evaluation system of the battery swap cabinet in the embodiment of the present invention includes:

[0088] The acquisition module 201 is used to deploy a distributed detection device network in the battery cluster of the battery swap cabinet and collect the battery original data set of the battery cluster of the battery swap cabinet;

[0089] The encoding module 202 is used to perform feature classification and vector encoding on the battery original data set to obtain multiple battery performance feature vectors;

[0090] Prediction module 203, used for inputting multiple battery performance feature vectors into a battery reuse potential prediction model set for prediction, and outputting a comprehensive battery reuse potential prediction result;

[0091] The calculation module 204 is used to set the evaluation node parameters based on the comprehensive battery reuse potential prediction results, and to perform adaptive sampling and dynamic optimization calculation on the evaluation node parameters to obtain a battery multi-stage evaluation control scheme.

[0092] Through the collaborative cooperation of the above-mentioned components, by deploying a distributed detection device network, and combining a multi-level data collection and processing mechanism, the present invention realizes the efficient acquisition and analysis of the battery life cycle data of the battery swap cabinet, and significantly improves the integrity and accuracy of data collection. An evaluation method combining deep learning with variational Bayesian inference is adopted, and the accuracy of battery performance evaluation is improved through multi-level feature extraction and dynamic optimization strategies. The cascade utilization decision-making mechanism based on multi-objective optimization, combined with scene feature matching and resource optimization allocation strategies, improves the matching accuracy of battery reuse. Through adaptive sampling and dynamic optimization control, the present invention can automatically adjust the evaluation strategy according to changes in battery performance, introduce a multi-layer variational Bayesian inference network, and combine the analysis of battery performance degradation laws to improve the reliability of the evaluation results. By integrating multiple prediction sub-models and a dynamic weight allocation mechanism, the present invention realizes a comprehensive evaluation of the potential for battery reuse, improves the prediction accuracy, and has a strong generalization ability.

[0093] above Figure 2 The battery reuse evaluation system for the battery swap cabinet in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The battery reuse evaluation device for the battery swap cabinet in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0094] Figure 3 It is a structural schematic diagram of a battery reuse evaluation device for a power swap cabinet provided in an embodiment of the present invention. The battery reuse evaluation device 300 for a power swap cabinet may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 310 (for example, one or more processors) and a memory 320, and one or more storage media 330 (for example, one or more massive storage device terminals) storing application programs 333 or data 332. Among them, the memory 320 and the storage medium 330 may be short-term storage or persistent storage. The program stored in the storage medium 330 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the battery reuse evaluation device 300 for the power swap cabinet. Furthermore, the processor 310 may be configured to communicate with the storage medium 330, and execute a series of instruction operations in the storage medium 330 on the battery reuse evaluation device 300 for the power swap cabinet, so as to implement the steps of the above-mentioned battery reuse evaluation method for the power swap cabinet.

[0095] The battery recycling evaluation device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input and output interfaces 360, and / or one or more operating systems 331, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will appreciate that Figure 3 The structure of the battery reuse evaluation device for the battery swap cabinet shown does not constitute a limitation on the battery reuse evaluation device for the battery swap cabinet provided by the present invention, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0096] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the battery reuse evaluation method for a battery swap cabinet.

[0097] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0098] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

[0099] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A battery recycling evaluation method for a battery swap cabinet, characterized in that: The method comprises: Deploy a distributed detection device network in the battery cluster of the battery swap cabinet, and collect the battery original data set of the battery cluster of the battery swap cabinet; Performing feature classification and vector encoding on the battery original data set to obtain multiple battery performance feature vectors; Inputting the plurality of battery performance feature vectors into a battery reuse potential prediction model set for prediction, and outputting a comprehensive battery reuse potential prediction result; The evaluation node parameters are set based on the comprehensive battery reuse potential prediction result, and the evaluation node parameters are adaptively sampled and dynamically optimized to obtain a multi-stage battery evaluation control scheme.

2. The battery recycling evaluation method for a battery swap cabinet according to claim 1 is characterized in that: The distributed detection device network is deployed in the battery cluster of the battery swap cabinet, and the original battery data set of the battery cluster of the battery swap cabinet is collected, including: The battery cluster of the battery swap cabinet is grouped and the distributed nodes are optimized according to the geographical location information and the use environment parameters to obtain a distributed detection device network; Performing data sampling cycle synchronization and time stamp alignment on the distributed detection device network to obtain synchronous sampling control parameters; Based on the synchronous sampling control parameters, high-frequency sampling is performed on the battery voltage parameter, current parameter and temperature parameter to obtain battery operating condition data, and time domain characteristic analysis and frequency domain characteristic analysis are performed on the battery operating condition data to obtain battery dynamic characteristic data; Constructing a voltage-current response curve based on the battery dynamic characteristic data, and calculating the battery internal resistance change value through the voltage-current response curve, and at the same time, correcting it in combination with the temperature compensation coefficient to obtain the battery internal resistance parameter; Record the charging and discharging process of the batteries in the battery cluster of the battery swap cabinet, calculate the charging and discharging capacity by the charge integration method, and count the number of cycles by combining the coulomb counting method to obtain the battery usage record data; The battery operating condition data, the battery dynamic characteristic data, the battery internal resistance parameter and the battery usage record data are time-series merged and data standardized to obtain a battery original data set.

3. The battery recycling evaluation method for a battery swap cabinet according to claim 2 is characterized in that: The battery original data set is subjected to feature classification and vector encoding to obtain a plurality of battery performance feature vectors, including: Based on the original data set of the battery, the voltage parameter, the current parameter and the temperature parameter are subjected to dimensionality reduction processing to obtain a battery parameter dimensionality reduction matrix, and a time series correlation analysis is performed on the battery parameter dimensionality reduction matrix to obtain a time series feature weight coefficient; Performing a weighted fusion operation on the time series feature weight coefficient and the battery parameter dimension reduction matrix to obtain a plurality of first battery performance features; Performing time series decomposition on the capacity data in the original data set of the battery, fitting the capacity decay curve by a double exponential function, and correcting it by combining the temperature stress coefficient to obtain the capacity decay model parameters; Building a battery capacity prediction function based on the capacity decay model parameters, calculating the capacity decay rate under different working conditions through Taylor expansion, and obtaining a capacity decay prediction value; The internal resistance data in the original data set of the battery are processed in segments, an internal resistance change function is established by a segmented linear regression method, and parameter optimization is performed in combination with historical data to obtain an internal resistance change model coefficient; Input the internal resistance change model coefficient into the prediction function, obtain the internal resistance prediction value at different time points through recursive calculation, and correct it with the temperature correction coefficient to obtain the internal resistance change prediction value; Performing feature vector quantization processing on the capacity decay prediction value and the internal resistance change prediction value to obtain a plurality of second battery performance characteristics; The plurality of first battery performance characteristics and the plurality of second battery performance characteristics are subjected to feature classification and vector encoding to obtain a plurality of battery performance characteristic vectors.

4. The battery recycling evaluation method for a battery swap cabinet according to claim 3 is characterized in that: The performing feature classification and vector encoding on the plurality of first battery performance features and the plurality of second battery performance features to obtain a plurality of battery performance feature vectors includes: Normalizing the plurality of first battery performance characteristics and the plurality of second battery performance characteristics, calculating an initial value of a feature space cluster center by a density peak detection algorithm, and obtaining a first cluster centroid; Inputting the first cluster centroid into the K-means algorithm, performing iterative calculations through the optimization objective function of minimizing the intra-class distance and maximizing the inter-class distance, to obtain the second cluster centroid; Performing a silhouette coefficient analysis on the second cluster centroid to obtain an optimal number of cluster categories, performing cluster division based on the optimal number of cluster categories, calculating the information entropy and weight coefficient of each feature by an entropy method, and obtaining a feature importance score; The feature importance score and the clustering result are weightedly combined, a hash coding strategy is generated through a local sensitive hash function, and a dimensionality reduction mapping is performed according to the hash coding strategy to obtain the feature association strength; Based on the feature association strength, high-order features are extracted and combined on the plurality of first battery performance features and the plurality of second battery performance features to obtain a plurality of battery performance feature vectors.

5. The battery recycling evaluation method for a battery swap cabinet according to claim 4 is characterized in that: The inputting the plurality of battery performance feature vectors into a battery reuse potential prediction model set for prediction, and outputting a comprehensive battery reuse potential prediction result, comprises: Performing subspace division based on the multiple battery performance feature vectors to obtain multiple independent feature subspaces; Inputting the multiple independent feature subspaces into a battery reuse potential prediction model set, the battery reuse potential prediction model set includes multiple battery reuse potential prediction sub-models, the battery reuse potential prediction sub-models include a feature extraction layer, a feature mapping layer and a prediction output layer, the feature extraction layer includes 3 fully connected layers, each fully connected layer uses BatchNormalization and ReLU activation function, the feature mapping layer includes 2 residual connection blocks, each residual connection block includes a two-layer perceptron and a jump connection, and the prediction output layer uses a Sigmoid activation function; Inputting the independent feature subspaces into the corresponding battery reuse potential prediction submodels respectively to obtain a submodel prediction result set, and dynamically assigning weights to the submodel prediction result set to obtain a fusion weight coefficient; The sub-model prediction result set is weighted analyzed based on the fusion weight coefficient to output a comprehensive battery reuse potential prediction result.

6. The battery recycling evaluation method for a battery swap cabinet according to claim 5 is characterized in that: The evaluation node parameters are set based on the comprehensive battery reuse potential prediction result, and the evaluation node parameters are adaptively sampled and dynamically optimized to obtain a battery multi-stage evaluation control scheme, including: Inputting the comprehensive battery reuse potential prediction result into the first layer variational Bayesian inference network for data distribution modeling, calculating the latent variable posterior distribution through the variational expectation maximization algorithm, and obtaining the battery performance distribution parameters; The battery performance distribution parameters are input into the second layer variational Bayesian inference network to perform battery performance evolution analysis, and the state transition probability matrix is ​​calculated in combination with random field theory to obtain the battery performance degradation law; Inputting the battery performance degradation law into the third-layer variational Bayesian inference network, performing probability modeling on the evaluation node parameters, and obtaining the evaluation node parameters; Performing multi-stage evaluation optimization based on the evaluation node parameters to obtain an evaluation optimization control strategy, and performing adaptive sampling on the evaluation optimization control strategy to obtain an adaptive sampling scheme; The adaptive sampling scheme is subjected to time series feature extraction to obtain an evaluation parameter dynamic adjustment matrix, and based on the evaluation parameter dynamic adjustment matrix, the evaluation accuracy and the evaluation efficiency are jointly optimized to obtain a battery multi-stage evaluation control scheme.

7. The battery recycling evaluation method for a battery swap cabinet according to claim 6 is characterized in that: The battery recycling assessment method for the battery swap cabinet also includes: The target layer is constructed for the multi-stage battery evaluation control scheme, and the three primary indicators of capacity decay rate, internal resistance change rate and cycle life are set by hierarchical analysis method to obtain an evaluation index system; A judgment matrix is ​​constructed according to the evaluation index system, the importance of the indicators is compared pairwise, and a matrix normalization operation is performed to obtain an indicator weight vector; Performing a consistency check on the indicator weight vector, solving the consistency ratio by an eigenvalue calculation method, and obtaining a target weight coefficient when the consistency ratio is less than a preset target value; Performing weighted calculation on battery performance based on the target weight coefficient to obtain a battery grading result, and performing scene feature matching based on the battery grading result to obtain a scene matching degree matrix; The optimal solution is searched through the scenario matching matrix to obtain the scenario priority ranking, and based on the scenario priority ranking, the constraint condition check is performed to ensure that the allocation plan meets the performance requirements and safety standards of each scenario, and the battery reuse execution plan is obtained.

8. A battery recycling evaluation system for a battery swap cabinet, characterized in that: The system is used to perform the battery reuse assessment method for a battery swap cabinet according to any one of claims 1 to 7, and comprises: A collection module, used to deploy a distributed detection device network in a battery cluster of a battery swap cabinet and collect the battery raw data set of the battery cluster of the battery swap cabinet; An encoding module, used for performing feature classification and vector encoding on the battery original data set to obtain a plurality of battery performance feature vectors; A prediction module, used for inputting the plurality of battery performance feature vectors into a battery reuse potential prediction model set for prediction, and outputting a comprehensive battery reuse potential prediction result; The calculation module is used to set the evaluation node parameters based on the comprehensive battery reuse potential prediction result, and to perform adaptive sampling and dynamic optimization calculation on the evaluation node parameters to obtain a battery multi-stage evaluation control scheme.

9. A battery recycling assessment device for a battery swap cabinet, characterized in that: The battery recycling assessment device for a power exchange cabinet includes: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the battery swap cabinet battery reuse evaluation device executes the battery swap cabinet battery reuse evaluation method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instruction is executed by the processor, the battery reuse evaluation method for the battery swap cabinet as described in any one of claims 1-7 is implemented.

Images