A power management device for energy storage equipment

By building an SOH value evaluation network in energy storage equipment, and performing correlation sorting and prediction of health factors based on the model of energy storage equipment and battery health status, the problem of inefficient evaluation of health factors in the existing technology is solved, and more accurate Battery health status assessment and pre-maintenance management.

CN118504426BActive Publication Date: 2024-11-19NANTONG INST OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410956110.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-17
Publication Date
2024-11-19
Estimated Expiration
2044-07-17

AI Technical Summary

Technical Problem

In the power consumption management of existing energy storage equipment, the evaluation of health factor is inefficient and insufficient objectivity, resulting in inaccurate evaluation of battery health status.

Method used

The battery service information is obtained through the battery service information acquisition module, and the initial health factors are sorted according to the battery health status SOH value of the energy storage device model, and the SOH value evaluation network is built, and the key health factors are predicted based on the number of charge and discharge cycles. , generate the predicted state values ​​of key health factors, and fit them through the SOH value evaluation network to generate k cycles of SOH values ​​to realize the pre-dimensional management of energy storage equipment.

Benefits of technology

It improves the efficiency and objectivity of health factor evaluation, improves the accuracy of battery health status evaluation results, and ensures the normal operation and service life of energy storage equipment.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118504426B_ABST
    Figure CN118504426B_ABST
Patent Text Reader

Abstract

The present invention discloses a power management device for energy storage equipment, which relates to the field of battery health management. The device includes: a battery service information acquisition module for obtaining battery service information; a key health factor generation module for generating key health factors; an SOH value evaluation network construction module for constructing an SOH value evaluation network; a key health factor prediction state value generation module for generating key health factor prediction state values; a k-cycle SOH value generation module for generating k-cycle SOH values; an energy storage device abnormal identification generation module for generating energy storage device abnormal identification; and a pre-maintenance management module for pre-maintenance management of energy storage equipment. The device solves the technical problems of low efficiency and insufficient objectivity of health factor evaluation in the power management of existing energy storage equipment, which leads to inaccurate battery health status evaluation results, and achieves the technical effect of improving the efficiency and objectivity of health factor evaluation and improving the accuracy of battery health status evaluation results.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of battery health management, and in particular to a power management device for energy storage equipment. Background Art

[0002] With the continuous development of energy storage technology, energy storage equipment is increasingly used in power systems, electric vehicles, renewable energy grid connection and other fields. The performance of energy storage equipment is directly related to the operating efficiency and safety of the entire system. Among them, the state of health (SOH) of the battery is one of the important indicators for evaluating the performance of energy storage equipment. The SOH value reflects the ability of the battery to store charge relative to a new battery, which is of great significance for predicting the remaining life of the battery, optimizing the charging and discharging strategy, and preventing equipment failure. In the existing energy storage equipment health assessment methods, the selection of health factors often relies on manual sorting. This method is mainly based on expert experience to analyze which health factors have a greater impact on the SOH value. However, due to the large differences in the degree of influence of health factors on the SOH value in different application scenarios, different models, and different charging and discharging modes, the use of manual sorting methods will lead to low evaluation efficiency and lack of objectivity.

[0003] In the current relevant technologies, the power management of energy storage equipment has technical problems such as low efficiency and lack of objectivity in health factor assessment, which leads to inaccurate battery health status assessment results. Summary of the invention

[0004] The present application provides an electricity management device for energy storage equipment, adopts the battery health status SOH value according to the energy storage equipment model, sorts the initial health factors by correlation, configures the timing prediction model based on the key health factors and the energy storage equipment model, builds the SOH value evaluation network, predicts the key health factors based on the number of charge and discharge cycles, generates the key health factor prediction status value, fits the key health factor prediction status value and the k-cycle label, generates the k-cycle SOH value and other technical means, so as to achieve the technical effect of improving the efficiency and objectivity of health factor evaluation and improving the accuracy of battery health status evaluation results.

[0005] The present application provides a power management device for energy storage equipment, comprising:

[0006] A battery service information acquisition module, the battery service information acquisition module is used to obtain battery service information, wherein the battery service information includes the number of charge and discharge cycles and the energy storage device model; a key health factor generation module, the key health factor generation module is used to sort the initial health factors by correlation according to the battery health status SOH value of the energy storage device model, and generate key health factors; an SOH value evaluation network construction module, the SOH value evaluation network construction module is used to configure a timing prediction model based on the key health factors and the energy storage device model, and construct an SOH value evaluation network; a key health factor prediction state value generation module, the key health factor prediction state value generation module is used to generate a key health factor prediction state value based on the number of charge and discharge cycles The key health factor is predicted by the number to generate a key health factor prediction state value, wherein the key health factor prediction state value has a k-cycle label; a k-cycle SOH value generation module, the k-cycle SOH value generation module is used to fit the key health factor prediction state value and the k-cycle label through the SOH value evaluation network to generate a k-cycle SOH value; an energy storage device abnormal identification generation module, the energy storage device abnormal identification generation module is used to generate an energy storage device abnormal identification based on the k-cycle label when the k-cycle SOH value does not meet the k-cycle SOH threshold; a pre-maintenance management module, the pre-maintenance management module is used to perform pre-maintenance management of the energy storage device according to the energy storage device abnormal identification.

[0007] In a possible implementation, the key health factor generation module includes:

[0008] A retrieval unit, the retrieval unit is used to retrieve the battery health state SOH value and the initial health factor with the energy storage device model as a constraint, and generate a SOH retrieval value set and an initial health factor retrieval value set, wherein the initial health factor retrieval value set corresponds one-to-one to the SOH retrieval value set; a dimension reduction processing unit, the dimension reduction processing unit is used to perform dimension reduction processing on the SOH retrieval value set, construct a benchmark SOH value sequence, traverse the initial health factor retrieval value set to perform dimension reduction processing, and construct multiple comparison health factor value sequences; a gray correlation evaluation unit, the gray correlation evaluation unit is used to perform gray correlation evaluation based on the benchmark SOH value sequence and the multiple comparison health factor value sequences, and generate a correlation evaluation value set, wherein the correlation evaluation value set corresponds one-to-one to the initial health factor; a key health factor extraction unit, the key health factor extraction unit is used to extract the key health factors whose correlation evaluation values ​​are greater than or equal to the correlation evaluation threshold from the initial health factors based on the correlation evaluation value set.

[0009] In a possible implementation, the retrieval unit includes:

[0010] A retrieval subunit, the retrieval subunit is used to retrieve the battery health state SOH value and the initial health factor with the energy storage device model as a constraint, and obtain a first SOH retrieval value set and a first initial health factor retrieval value set; an aggregation subunit, the aggregation subunit is used to perform aggregation analysis on the first initial health factor retrieval value set to generate multiple clusters of first initial health factor retrieval values, wherein the multiple clusters of first initial health factor retrieval values ​​have a centroid initial health factor retrieval value; a grouping subunit, the grouping subunit is used to group the first SOH retrieval value set according to the multiple clusters of first initial health factor retrieval values ​​to generate multiple clusters of first SOH retrieval values; a de-extreme value fitting subunit, the de-extreme value fitting subunit is used to traverse the multiple clusters of first SOH retrieval values ​​to perform de-extreme value fitting and generate SOH fitting values; a retrieval value set generation subunit, the retrieval value set generation subunit is used to set the centroid initial health factor retrieval value to the initial health factor retrieval value set, and set the SOH fitting value to the SOH retrieval value set.

[0011] In a possible implementation, the aggregation subunit includes:

[0012] an initial health factor retrieval value acquisition microunit, the initial health factor retrieval value acquisition microunit is used to obtain a first group of initial health factor retrieval values ​​of the first initial health factor retrieval value set, until the Nth group of initial health factor retrieval values ​​of the first initial health factor retrieval value set is obtained; a normalization processing microunit, the normalization processing microunit is used to normalize the first group of initial health factor retrieval values ​​to construct a first coordinate, until the Nth group of initial health factor retrieval values ​​are normalized to construct an Nth coordinate; a cluster analysis microunit, the cluster analysis microunit is used to normalize the first coordinate until the Nth group of initial health factor retrieval values ​​are normalized to construct an Nth coordinate; The Nth coordinate, cluster analysis is performed based on the Euclidean distance threshold to generate a multi-cluster coordinate clustering result; a centroid sorting microunit, the centroid sorting microunit is used to traverse the multi-cluster coordinate clustering result to perform centroid sorting, and generate multiple centroid coordinates; a set grouping microunit, the set grouping microunit is used to group the first initial health factor retrieval value set according to the multi-cluster coordinate clustering result to generate the multi-cluster first initial health factor retrieval value; a set sorting microunit, the set sorting microunit is used to sort the first initial health factor retrieval value set according to the multiple centroid coordinates to generate the centroid initial health factor retrieval value.

[0013] In a possible implementation, the centroid sorting microunit includes:

[0014] A coordinate acquisition sub-micro-unit, the coordinate acquisition sub-micro-unit is used to obtain the first cluster coordinates of the first cluster coordinate clustering result until the Mth cluster coordinates; a coordinate set acquisition sub-micro-unit, the coordinate set acquisition sub-micro-unit is used to collect the first reference coordinate set with the first cluster coordinates as the reference point and the preset number of neighboring coordinates as the constraint, until the Mth neighboring coordinate set is collected with the Mth cluster coordinates as the reference point and the preset number of neighboring coordinates as the constraint, to obtain a second reference coordinate set with the first cluster coordinates as the neighboring coordinates; a centroid acquisition sub-micro-unit, the centroid acquisition sub-micro-unit is used to calculate the inverse of the mean of the distances from the first reference coordinate set and the second reference coordinate set to the first cluster coordinates, and set it as the first centroid, until the Mth centroid is obtained; a centroid coordinate generation sub-micro-unit, the centroid coordinate generation sub-micro-unit is used to set the coordinates corresponding to the maximum values ​​of the first centroid up to the Mth centroid as the first cluster centroid coordinates, and add them to the multiple centroid coordinates.

[0015] In a possible implementation, the key health factor generation module includes:

[0016] An initial health factor construction unit, wherein the initial health factor construction unit is used for the initial health factor to include at least trickle current rise time, constant current rise time, constant voltage drop time, cycle end voltage recovery value, battery charging temperature, and battery discharging temperature.

[0017] In a possible implementation, the SOH value evaluation network construction module includes:

[0018] A networked charging history curve sequence acquisition unit, the networked charging history curve sequence acquisition unit is used to collect the charging history curve sequence based on the key health factor and the energy storage device model as constraints, wherein the charging history curve sequence corresponds to the number of charge and discharge cycles one by one; a convolutional feature extractor construction unit, the convolutional feature extractor construction unit is used to construct a convolutional feature extractor, wherein the convolutional feature extractor is used to process the charging history curve sequence to extract the key health factor time series characteristics of the key health factor; a feature extraction unit, the feature extraction unit is used to extract the features of the charging history curve sequence through the convolutional feature extractor to obtain the key health factor time series feature record data, wherein the number of charge and discharge cycles has a SOH true value label sequence; a splitting unit, the splitting unit is used to process the key health factor The key health factor time series feature record data is split to obtain a first group of key health factor feature sequences, wherein the first group of key health factor feature sequences has a sequence endpoint cycle number label, and the first group of key health factor feature sequences has at least three sequence nodes; a SOH true value label sequence sorting unit, the SOH true value label sequence sorting unit is used to sort the SOH true value label sequence according to the first group of key health factor feature sequences to obtain the sequence endpoint SOH true value label; a training data set construction unit, the training data set construction unit is used to use the first group of key health factor feature sequences and the sequence endpoint cycle number label as the input of the long short-term memory neural network, use the sequence endpoint SOH true value label as the supervision of the long short-term memory neural network, construct a training data set, and configure the SOH value evaluation network.

[0019] In a possible implementation, the key health factor prediction state value generation module includes:

[0020] A local charging history curve sequence acquisition unit, the local charging history curve sequence acquisition unit is used to acquire a local charging history curve sequence according to the number of charge and discharge cycles; a key health factor fluctuation curve construction unit, the key health factor fluctuation curve construction unit is used to construct a key health factor fluctuation curve according to the local charging history curve sequence; a segmentation unit, the segmentation unit is used to segment the key health factor fluctuation curve according to the charging mode, and obtain the key health factor fluctuation curve segmentation result; a health factor average change speed extraction unit, the health factor average change speed extraction unit is used to traverse the key health factor fluctuation curve segmentation result, and extract the health factor average change speed, wherein the health factor average change speed corresponds to the charging mode one by one; A cycle label configuration unit, the cycle label configuration unit is used to configure the k-time cycle label through the user end; a predicted state value generation unit, the predicted state value generation unit is used to predict the key health factor based on the k-time cycle label and the average change rate of the first health factor in the charging mode with the maximum SOH value loss, and generate a first predicted state value, and predict the key health factor based on the k-time cycle label and the average change rate of the second health factor in the charging mode with the minimum SOH value loss, and generate a second predicted state value; a predicted state interval construction unit, the predicted state interval construction unit is used to construct a predicted state interval according to the first predicted state value and the second predicted state value, and add the predicted state interval to the predicted state value of the key health factor.

[0021] It is intended to obtain battery service information through a power management device for energy storage equipment proposed in this application through a battery service information acquisition module, wherein the battery service information includes the number of charge and discharge cycles and the model of the energy storage equipment, and to sort the initial health factors according to the battery health status SOH value of the energy storage equipment model through the key health factor generation module to generate key health factors, and to configure a time series prediction model based on the key health factors and the energy storage equipment model through the SOH value evaluation network construction module to construct an SOH value evaluation network, and to predict the key health factors based on the number of charge and discharge cycles through the key health factor prediction state value generation module to generate key health factors. A key health factor prediction state value, wherein the key health factor prediction state value has a k-cycle label, and a k-cycle SOH value generation module is used to fit the key health factor prediction state value and the k-cycle label through an SOH value evaluation network to generate a k-cycle SOH value, and when the k-cycle SOH value does not meet the k-cycle SOH threshold through an energy storage device abnormal identification generation module, an energy storage device abnormal identification is generated based on the k-cycle label, and a pre-maintenance management module is used to perform pre-maintenance management of the energy storage device according to the energy storage device abnormal identification, thereby achieving the technical effect of improving the efficiency and objectivity of health factor evaluation and improving the accuracy of battery health status evaluation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the device according to the embodiment of the present application. It should be understood that the previous or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0023] Figure 1 A schematic diagram of the structure of a power management device for energy storage equipment provided in an embodiment of the present application.

[0024] Figure 2 A structural schematic diagram of a key health factor prediction state value generation module of a power consumption management device for energy storage equipment provided in an embodiment of the present application.

[0025] Explanation of the reference numerals: battery service information acquisition module 10, key health factor generation module 20, SOH value evaluation network construction module 30, key health factor prediction state value generation module 40, k-cycle SOH value generation module 50, energy storage device abnormality identification generation module 60, pre-maintenance management module 70. DETAILED DESCRIPTION

[0026] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0027] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0028] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, device, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0029] The present application embodiment provides a power management device for energy storage equipment, such as Figure 1 As shown, the device comprises:

[0030] The battery service information acquisition module 10 is used to obtain the battery service information, wherein the battery service information includes the number of charge and discharge cycles and the energy storage device model. Specifically, the battery service information refers to various data generated by the battery during the service process, such as the number of charge and discharge cycles, temperature, voltage, etc., which reflects the performance status and health status of the battery. Specifically, the battery service information acquisition module 10 is activated, and the battery service information acquisition module 10 is connected to the battery management system (BMS, a system for managing and controlling the battery status, which can monitor the battery voltage, current, temperature and other parameters in real time) of the energy storage device or other related databases, and the number of charge and discharge cycles of the battery is extracted from the BMS or database, wherein the battery is fully charged, discharged, and then fully charged as a cycle, and the accumulated number of cycles is an important indicator for evaluating the degree of battery aging. As the number of cycles increases, the performance of the battery will gradually decline, and the number of charge and discharge cycles indicates how many complete charge and discharge processes the battery has experienced since its brand new state. At the same time, the battery service information acquisition module 10 acquires the model information of the energy storage device from the data source, wherein the energy storage device has different models and specifications according to the application scenario and design requirements, and different models correspond to different battery configurations and performance parameters.

[0031] The key health factor generation module 20 is used to sort the initial health factors by correlation according to the battery health status SOH value of the energy storage device model to generate key health factors. Specifically, the key health factor generation module 20 receives the energy storage device model information transmitted from the battery service information acquisition module 10, and retrieves or requests the historical battery health status (SOH) data and initial health factor data related to the model from the database according to the energy storage device model, wherein the battery health status SOH value represents the ratio of the current performance of the battery to its performance in a brand new state, and is an important indicator for evaluating battery performance, which can be obtained through laboratory testing or a prediction model based on historical data; the initial health factor refers to all possible factors or indicators related to the battery health status when evaluating the battery health status, including the number of battery charge and discharge times, internal resistance, capacity decay rate, temperature, etc. The acquired initial health factor data is cleaned to remove outliers, missing values ​​or unreasonable data, and the data is standardized or normalized to ensure that the initial health factors of different dimensions have the same weight in the analysis. The correlation between the initial health factor and the battery health state SOH value is analyzed by using statistical methods (such as Pearson correlation coefficient, Spearman rank correlation coefficient, etc.) or feature selection techniques in machine learning (such as model-based feature selection, tree-based feature importance, etc.), and the contribution of each initial health factor to the prediction of the battery health state SOH value is evaluated to determine its importance. Based on the results of the correlation analysis, a suitable threshold or selection strategy is set to screen out the key health factors that are strongly correlated with the battery health state SOH value, that is, the key health factor is the initial health factor that has a significant impact on the battery health state SOH value when evaluating the battery health state, and can accurately reflect the health status of the battery.

[0032] In a possible implementation, the key health factor generation module 20 includes:

[0033] A retrieval unit, the retrieval unit is used to retrieve the battery health state SOH value and the initial health factor with the energy storage device model as a constraint, and generate an SOH retrieval value set and an initial health factor retrieval value set, wherein the initial health factor retrieval value set corresponds to the SOH retrieval value set one by one. Specifically, the retrieval unit receives the specified energy storage device model as a retrieval constraint, retrieves the battery health state SOH value and initial health factor data corresponding to the model in the database with the energy storage device model as an index, arranges the retrieved battery health state SOH value in order of time or number of cycles, etc., to form an SOH retrieval value set; at the same time, arranges the corresponding initial health factor data in the same order to form an initial health factor retrieval value set, so that the elements of the two sets correspond one by one. A de-dimensionalization processing unit, the de-dimensionalization processing unit is used to perform de-dimensionalization processing on the SOH retrieval value set, construct a benchmark SOH value sequence, traverse the initial health factor retrieval value set to perform de-dimensionalization processing, and construct multiple comparison health factor value sequences. Specifically, each SOH search value in the SOH search value set is de-dimensionalized (such as normalized, standardized, etc.) to eliminate the unit or dimension differences between different SOH search values, and to construct a benchmark SOH value sequence; each initial health factor value in the initial health factor search value set is traversed and de-dimensionalized to generate multiple comparison health factor value sequences, each comparison health factor value sequence corresponds to an initial health factor. A gray correlation evaluation unit is used to perform gray correlation evaluation based on the benchmark SOH value sequence and the multiple comparison health factor value sequences to generate a correlation evaluation value set, wherein the correlation evaluation value set corresponds to the initial health factor one by one. Specifically, based on the benchmark SOH value sequence and each comparison health factor value sequence, the gray correlation between the two is calculated. The gray correlation is an indicator that measures the similarity or degree of association between two sequences. The calculated gray correlation between each initial health factor and the SOH value is used as the correlation evaluation value to form a correlation evaluation value set, and each element in the correlation evaluation value set corresponds to an initial health factor one by one. A key health factor extraction unit is used to extract the key health factors whose correlation evaluation values ​​are greater than or equal to the correlation evaluation threshold from the initial health factors based on the correlation evaluation value set. Specifically, according to actual needs or expert advice, a correlation evaluation threshold is set, and the correlation evaluation value set is traversed to find the initial health factors whose correlation evaluation values ​​are greater than or equal to the correlation evaluation threshold. These factors are factors with a high correlation with the SOH value and have a greater impact on the health status of the energy storage device, namely, key health factors.This implementation method uses grey correlation to sort the correlation between the battery health status SOH value and the initial health factor. The grey correlation evaluation can solve the uncertainty problem of influencing factors between multiple indicators, achieving the technical effect of improving the accuracy of correlation evaluation.

[0034] In a possible implementation, the retrieval unit includes:

[0035] A retrieval subunit, the retrieval subunit is used to retrieve the battery state of health SOH value and the initial health factor with the energy storage device model as a constraint, and obtain a first SOH retrieval value set and a first initial health factor retrieval value set. Specifically, the retrieval subunit receives the input energy storage device model as a constraint condition, retrieves the battery state of health SOH value and initial health factor data matching the specified model in the database, arranges the SOH values ​​retrieved for the first time in order of time or number of cycles, etc., to form a first SOH retrieval value set, and at the same time, arranges the corresponding initial health factor data retrieved for the first time in the same order to form a first initial health factor retrieval value set. An aggregation subunit, the aggregation subunit is used to perform aggregation analysis on the first initial health factor retrieval value set to generate multiple clusters of first initial health factor retrieval values, wherein the multiple clusters of first initial health factor retrieval values ​​have a centroid initial health factor retrieval value. Specifically, a cluster analysis (such as K-means clustering) is performed on the first initial health factor retrieval value set, and according to the clustering results, the initial health factor data is divided into multiple clusters, each cluster having a centroid (i.e., the average value or center point of all the data in the cluster). A grouping subunit, the grouping subunit is used to group the first SOH retrieval value set according to the multiple clusters of first initial health factor retrieval values ​​to generate multiple clusters of first SOH retrieval values. Specifically, the first SOH retrieval value set is grouped according to the multiple clusters of first initial health factor retrieval values, and the principle of grouping is to ensure that the SOH values ​​corresponding to the initial health factor retrieval values ​​of the same cluster are also divided into the same cluster, so that the SOH retrieval values ​​of each cluster correspond to a centroid initial health factor retrieval value. A de-extreme value fitting subunit, the de-extreme value fitting subunit is used to traverse the multiple clusters of first SOH retrieval values ​​to perform de-extreme value fitting and generate SOH fitting values. Specifically, traverse each cluster of the first SOH retrieval values ​​of multiple clusters, perform de-extreme processing on the SOH retrieval values ​​of each cluster (such as removing the maximum and minimum values, or determining and removing the extreme values ​​according to statistical methods), and fit the de-extreme SOH retrieval values ​​(such as calculating the mean, median, or mode value, etc.) to generate SOH fitting values. The retrieval value set generation subunit is used to set the centroid initial health factor retrieval value to the initial health factor retrieval value set, and set the SOH fitting value to the SOH retrieval value set. Specifically, the centroid initial health factor retrieval value generated by the aggregation subunit is set as the final initial health factor retrieval value set, and the SOH fitting value generated by the de-extreme fitting subunit is set as the final SOH retrieval value set. This implementation method provides a high-quality data basis for grey correlation evaluation by preprocessing and organizing the initially retrieved data, thereby achieving the technical effect of improving the accuracy and efficiency of subsequent analysis.

[0036] In a possible implementation, the aggregation subunit includes:

[0037] An initial health factor retrieval value acquisition micro-unit, the initial health factor retrieval value acquisition micro-unit is used to obtain the first group of initial health factor retrieval values ​​of the first initial health factor retrieval value set, until the Nth group of initial health factor retrieval values ​​of the first initial health factor retrieval value set is obtained. Specifically, the first initial health factor retrieval value set is read, and the first group of initial health factor retrieval values ​​in the first initial health factor retrieval value set sorted by time or number of cycles are sequentially obtained, until the Nth group of initial health factor retrieval values. A normalization processing micro-unit, the normalization processing micro-unit is used to normalize the first group of initial health factor retrieval values, construct a first coordinate, until the Nth group of initial health factor retrieval values ​​are normalized and the Nth coordinate is constructed. Specifically, each initial health factor value in the first group of initial health factor retrieval values ​​is normalized, and each initial health factor value is converted to a value between 0 and 1, or converted to a value with the same dimension. After the first group of initial health factor retrieval values ​​are normalized, the first coordinate is constructed, and the above steps are repeated until the Nth group of initial health factor retrieval values ​​are normalized and the Nth coordinate is constructed. A cluster analysis micro-unit, the cluster analysis micro-unit is used to perform cluster analysis on the first coordinate to the Nth coordinate based on the Euclidean distance threshold, and generate a multi-cluster coordinate clustering result. Specifically, a Euclidean distance threshold is set as the basis for determining whether the data points belong to the same cluster during cluster analysis, and cluster analysis is performed on the first coordinate to the Nth coordinate, and the coordinate points are divided into multiple clusters based on the Euclidean distance between the coordinate points. During the clustering process, if the Euclidean distance between two coordinate points is less than the set Euclidean distance threshold, they are divided into the same cluster. A centroid sorting micro-unit, the centroid sorting micro-unit is used to traverse the multi-cluster coordinate clustering results to perform centroid sorting and generate multiple centroid coordinates. Specifically, each coordinate cluster is traversed, the average value (centroid) of all coordinate points in the cluster is calculated, and the calculated centroid coordinate is used as the representative point of the cluster. After the traversal is completed, multiple centroid coordinates are generated. A set grouping micro-unit, the set grouping micro-unit is used to group the first initial health factor retrieval value set according to the multi-cluster coordinate clustering result, and generate the multi-cluster first initial health factor retrieval value. Specifically, according to the multi-cluster coordinate clustering result, each group of initial health factor retrieval values ​​in the first initial health factor retrieval value set is divided into corresponding clusters. After the division is completed, multiple clusters of first initial health factor retrieval values ​​are generated, and each cluster represents an initial health factor pattern. A set sorting micro-unit is used to sort the first initial health factor retrieval value set according to the multiple centroid coordinates to generate the centroid initial health factor retrieval value. Specifically, using the generated multiple centroid coordinates as a reference, the initial health factor retrieval value corresponding to or closest to the centroid coordinate is selected from the first initial health factor retrieval value set as the centroid initial health factor retrieval value.This implementation eliminates the influence of scale through normalization processing, allowing the clustering algorithm to more accurately calculate the distance and similarity between data points, achieving the technical effect of improving the accuracy of clustering analysis.

[0038] In a possible implementation, the centroid sorting microunit includes:

[0039] A coordinate acquisition sub-micro-unit, the coordinate acquisition sub-micro-unit is used to obtain the first cluster coordinates of the first cluster coordinate clustering result until the Mth cluster coordinates. Specifically, each coordinate clustering result (i.e., each cluster) is traversed, and all coordinates in each cluster are obtained. For example, starting from the first cluster, all coordinates in the first cluster are collected, and then the second cluster, the third cluster, and the Mth cluster are continued. A coordinate set acquisition sub-micro-unit, the coordinate set acquisition sub-micro-unit is used to collect the first reference coordinate set with the first cluster coordinate as the reference point and the preset number of neighboring coordinates as the constraint, until the Mth cluster coordinate is used as the reference point and the Mth neighboring coordinate set is collected with the preset number of neighboring coordinates as the constraint, and the second reference coordinate set with the first cluster coordinate as the neighboring coordinate is obtained. Specifically, for each cluster coordinate (as a reference point), the neighboring coordinates around the coordinate are collected, and the number of collected coordinates is determined by the preset number of neighboring coordinates. The collected coordinates constitute a set, i.e., the first reference coordinate set. For example, for a coordinate in the first cluster, a preset number of coordinates near the coordinate are found to form a reference coordinate set, and this process is repeated until a corresponding neighboring coordinate set is found for each coordinate in each cluster. For each reference point in the first cluster, check whether it falls within the proximity of other clusters. If so, obtain a second reference coordinate set with the coordinates in the first cluster as the neighboring coordinates. The centroid acquisition sub-micro unit is used to calculate the reciprocal of the mean distance between the first reference coordinate set and the second reference coordinate set and the coordinates in the first cluster, and set it as the first centroid until the Mth centroid is obtained. Specifically, the reciprocal of the mean distance between each first reference coordinate set and the corresponding second reference coordinate set to the reference point (coordinates in the cluster) is calculated. This reciprocal is called the centroid, which is used to measure the suitability of the coordinate as the centroid. For each cluster, a centroid set is obtained, representing the possibility of each coordinate in the cluster being the centroid. The centroid coordinate generation sub-micro unit is used to set the coordinates corresponding to the maximum values ​​of the first centroid up to the Mth centroid as the centroid coordinates of the first cluster, and add them to the multiple centroid coordinates. Specifically, the centroid set of each cluster is traversed to find the maximum centroid value, and the coordinates corresponding to the value are set as the centroid coordinates of the cluster. The centroid coordinates of all clusters are added to a set to form multiple centroid coordinates. This implementation method obtains a representative and reliable centroid coordinate set through the steps of coordinate acquisition, coordinate set collection, centroid degree calculation and centroid coordinate generation, achieving the technical effect of accurately and efficiently determining the centroid coordinates of each cluster.

[0040] In a possible implementation, the key health factor generation module 20 includes:

[0041] The initial health factor construction unit is used for the initial health factor, which includes at least the trickle rise time, the constant current rise time, the constant voltage drop time, the cycle end voltage recovery value, the battery charging temperature, and the battery discharge temperature. Specifically, in the early stage of battery charging, in order to activate the chemical substances inside the battery, a smaller current is used for charging. This period of time is called the trickle rise time; after trickle charging, the battery enters the constant current charging stage, at which time the charging current remains unchanged until the battery voltage reaches a preset value, and this period of time is called the constant current rise time; after the battery voltage reaches the preset value, the charger will switch to the constant voltage charging mode, at which time the battery voltage remains unchanged, and the charging current gradually decreases until the charging is completed, and this period of time is called the constant voltage drop time; after the discharge is completed, the battery voltage will have a short recovery process, and this recovered voltage value is called the cycle end voltage recovery value; the battery charging temperature and the battery discharge temperature refer to the temperature of the battery during the charging and discharging process, respectively. The parameters in this implementation method reflect the activity of the chemical reaction inside the battery, the stability of the electrode material, and the thermal stability of the battery, etc., and achieve the technical effect of accurately constructing the initial health factor for evaluating the health status of the battery.

[0042] The SOH value evaluation network construction module 30 is used to configure the timing prediction model based on the key health factor and the energy storage device model to construct the SOH value evaluation network. Specifically, the SOH value evaluation network construction module 30 receives the key health factor transmitted from the key health factor generation module 20, and receives the energy storage device model information transmitted from the battery service information acquisition module 10. According to the application scenario and data characteristics, a timing prediction model (a prediction model that can process time series data and is used to analyze data that changes over time) is selected as the basis of the SOH value evaluation network, wherein the timing prediction model can select a recurrent neural network (RNN), a long short-term memory network (LSTM), etc. According to the energy storage device model, the parameters and architecture of the timing prediction model are configured. Different energy storage device models correspond to different battery characteristics and performance parameters, so it is necessary to adjust the parameters and architecture of the timing prediction model according to the actual situation. Further data preprocessing is performed on the key health factors, such as feature scaling, encoding, etc., to meet the input requirements of the time series prediction model. If necessary, the data can be grouped or labeled according to the energy storage device model so that the differences between different energy storage device models can be considered during the time series prediction model training process. The time series prediction model is trained using historical data, which includes key health factors and corresponding SOH values, so that the time series prediction model learns the mapping relationship between key health factors and SOH values. The time series prediction model is repeatedly trained and verified until satisfactory performance is achieved, and the SOH value evaluation network is obtained, that is, the SOH value evaluation network is a neural network based on the time series prediction model, which is used to evaluate the SOH value of the battery according to the key health factors and the energy storage device model.

[0043] In a possible implementation, the SOH value evaluation network construction module 30 includes:

[0044] A networked charging history curve sequence acquisition unit, the networked charging history curve sequence acquisition unit is used to collect the charging history curve sequence online with the key health factor and the energy storage device model as constraints, wherein the charging history curve sequence corresponds to the number of charge and discharge cycles. Specifically, the unit is connected to the corresponding data warehouse or charging management system through the network, and the historical charging data is retrieved with the key health factors (such as trickle rise time, cross current rise time, etc.) and the energy storage device model as constraints, and the retrieved historical charging data is arranged in chronological order to form a charging history curve sequence, each charging history curve sequence represents a complete charge and discharge cycle, and the charging history curve sequence records the changes in the key health factors of the energy storage device in multiple charge and discharge cycles. A convolution feature extractor construction unit, the convolution feature extractor construction unit is used to construct a convolution feature extractor, wherein the convolution feature extractor is used to process the charging history curve sequence to extract the key health factor time series characteristics of the key health factor. Specifically, a convolutional neural network architecture is designed to extract the time series features of key health factors from the charging history curve sequence, that is, the convolutional feature extractor is a feature extractor based on the convolutional neural network, which is used to learn and extract useful features from time series data. A feature extraction unit, the feature extraction unit is used to extract features from the charging history curve sequence through the convolutional feature extractor to obtain key health factor time series feature record data, wherein the number of charge and discharge cycles has a SOH true value label sequence; a splitting unit, the splitting unit is used to split the key health factor time series feature record data to obtain a first group of key health factor feature sequences, wherein the first group of key health factor feature sequences has a sequence end cycle number label, and the first group of key health factor feature sequences has at least three sequence nodes. Specifically, according to time or charge and discharge cycle number, the extracted key health factor time series feature record data is split into multiple shorter key health factor feature sequences, each key health factor feature sequence has a clear start and end point, that is, a start and end cycle number label, and each key health factor feature sequence contains at least three sequence nodes to ensure that sufficient time series information can be captured. The SOH true value label sequence sorting unit is used to sort the SOH true value label sequence according to the first group of key health factor feature sequences to obtain the SOH true value label at the end of the sequence. Specifically, according to the number of sequence end cycles of the key health factor feature sequence obtained after splitting, the corresponding label value is selected from the SOH true value label sequence, and the SOH true value label is the SOH value of the energy storage device obtained by actual testing or experiment at a specific number of cycles, which serves as the standard answer (supervisory signal) for model training.A training data set construction unit, the training data set construction unit is used to use the first group of key health factor feature sequences and the sequence endpoint cycle number labels as inputs of the long short-term memory neural network, use the sequence endpoint SOH true value labels as supervision of the long short-term memory neural network, construct a training data set, and configure the SOH value evaluation network. Specifically, the key health factor feature sequence obtained by splitting and the corresponding sequence endpoint cycle number labels are used as inputs of the long short-term memory neural network (LSTM), and the sequence endpoint SOH true value labels are used as supervision signals of the LSTM. These data are used to construct a training data set, configure the SOH value evaluation network, and train the network so that it can learn the mapping relationship from the key health factor feature sequence to the SOH value. This implementation method uses a convolutional feature extractor to extract the key health factor time series features from a complex charging history curve sequence, provides effective input information for the model, and reduces the complexity of the model by splitting the long time series feature sequence into shorter sequences, improves the training efficiency, and captures more time series information, achieving the technical effect of improving the accuracy of the SOH value evaluation network in evaluating the SOH value of energy storage equipment.

[0045] A key health factor prediction state value generation module 40 is used to predict the key health factor based on the number of charge and discharge cycles to generate a key health factor prediction state value, wherein the key health factor prediction state value has a k-cycle label. Specifically, the key health factor prediction state value generation module 40 receives the key health factor transmitted from the key health factor generation module 20 and the number of charge and discharge cycles of the energy storage device transmitted by the battery service information acquisition module 10. The historical data of the key health factor is obtained, including the corresponding relationship between the key health factor and the number of charge and discharge cycles, and the historical data is used as training data. A prediction model is selected, such as a linear regression model, a support vector machine (SVM), a random forest or a neural network, etc., for predicting the state value of the key health factor according to the number of charge and discharge cycles. The prediction model is trained using historical data, and through training, the prediction model learns the change law of the key health factor with the number of charge and discharge cycles. The number of charge and discharge cycles of the current energy storage device is input into the trained prediction model. The prediction model predicts the state value of the corresponding key health factor based on the input number of cycles, and adds a k-cycle label to the predicted key health factor state value, indicating that the predicted state value of the key health factor is the prediction result after k cycles.

[0046] like Figure 2 As shown, in a possible implementation, the key health factor prediction state value generating module 40 includes:

[0047] A local charging history curve sequence acquisition unit, the local charging history curve sequence acquisition unit is used to acquire a local charging history curve sequence according to the number of charge and discharge cycles. Specifically, the charging and discharging records of the energy storage device are read to determine the number of charge and discharge cycles. According to the number of charge and discharge cycles, the corresponding charging history data is retrieved from the local database, and the retrieved history data is arranged in chronological order to form a local charging history curve sequence, the local charging history curve sequence is a curve sequence that records the data changes of the key health factors of the energy storage device in multiple charge and discharge cycles, and changes with time. A key health factor fluctuation curve construction unit, the key health factor fluctuation curve construction unit is used to construct a key health factor fluctuation curve according to the local charging history curve sequence. Specifically, the data points of the key health factors are extracted from the local charging history curve sequence, and these data points are plotted into a curve graph in chronological order to form a key health factor fluctuation curve, which reflects the fluctuation of the key health factors with the number of charge and discharge cycles. A segmentation unit, the segmentation unit is used to segment the key health factor fluctuation curve according to the charging mode to obtain the segmentation result of the key health factor fluctuation curve. Specifically, different charging modes in the charging history are identified (specific modes or stages in which parameters such as current and voltage change over time during the charging process, such as trickle charging, constant current charging, constant voltage charging, etc.), and the key health factor fluctuation curve is segmented according to the charging mode. A health factor average change rate extraction unit is used to traverse the segmentation results of the key health factor fluctuation curve and extract the average change rate of the health factor, wherein the average change rate of the health factor corresponds to the charging mode one by one. Specifically, the key health factor fluctuation curve segment under each charging mode is calculated, and the average change rate of the key health factor in each curve segment is calculated, that is, the average rate at which the key health factor changes with the number of cycles under a specific charging mode. A cycle label configuration unit is used to configure the k-cycle label through the user end. Specifically, on the user end interface, an input box or selection box is provided for the user to input or select the k value. The user configures the k-cycle label according to actual needs. The k-cycle label is a specific cycle number label specified by the user for predicting the future state of the key health factor. A predicted state value generating unit, wherein the predicted state value generating unit is used to predict the key health factor based on the k-cycle labels and the average change rate of the first health factor in the charging mode with the maximum SOH value loss, and generate a first predicted state value; based on the k-cycle labels, the key health factor is predicted based on the average change rate of the second health factor in the charging mode with the minimum SOH value loss, and generate a second predicted state value.Specifically, the charging mode with the largest SOH value loss is identified, the average change rate of the first health factor under the mode is extracted, and the state value (first predicted state value) of the key health factor after k cycles is predicted based on the k-cycle label and the average change rate of the first health factor; the charging mode with the smallest SOH value loss is identified, the average change rate of the second health factor under the mode is extracted, and the state value (second predicted state value) of the key health factor after k cycles is predicted based on the k-cycle label and the average change rate of the second health factor. A prediction state interval construction unit is used to construct a prediction state interval according to the first prediction state value and the second prediction state value, and add the prediction state interval to the prediction state value of the key health factor. Specifically, the first prediction state value and the second prediction state value are used as the two endpoints of the interval to construct a prediction state interval containing these two endpoints. This implementation method predicts based on the charging modes with the largest and smallest SOH value losses, constructs a prediction state interval, and achieves the technical effect of providing comprehensive prediction information.

[0048] The k-cycle SOH value generation module 50 is used to fit the key health factor prediction state value and the k-cycle label through the SOH value evaluation network to generate the k-cycle SOH value. Specifically, the k-cycle SOH value generation module 50 receives the key health factor prediction state value and the corresponding k-cycle label from the key health factor prediction state value generation module 40, and obtains the SOH value evaluation network constructed by the SOH value evaluation network construction module 30. The key health factor prediction state value and the k-cycle label are integrated as input data of the SOH value evaluation network, and the integrated data is input into the SOH value evaluation network. The SOH value evaluation network fits the input key health factor prediction state value and the k-cycle label according to previous learning and training, and generates the corresponding k-cycle SOH value through the time series prediction model. The k-cycle SOH value represents the health state SOH value of the battery after k cycles, which is used to evaluate the performance and remaining life of the battery after a specific number of cycles.

[0049] The energy storage device abnormal identification generation module 60 is used to generate an energy storage device abnormal identification based on the k-cycle label when the k-cycle SOH value does not meet the k-cycle SOH threshold. Specifically, the energy storage device abnormal identification generation module 60 receives the k-cycle SOH value from the k-cycle SOH value generation module 50, obtains or sets the corresponding k-cycle SOH threshold, and the k-cycle SOH threshold is set according to the performance standard, historical data or expert advice of the energy storage device, and is used to determine whether the k-cycle SOH value is abnormal. The received k-cycle SOH value is compared with the set k-cycle SOH threshold to determine whether the k-cycle SOH value is lower than the k-cycle SOH threshold. If so, it is considered that the health status of the energy storage device has problems after k cycles or is about to reach its service life. At this time, the energy storage device abnormal identification is generated according to the k-cycle label, including the unique identifier of the energy storage device, the abnormality type (such as the SOH value is too low), the time when the abnormality occurred (known through the k-cycle label), and other information.

[0050] Pre-maintenance management module 70, the pre-maintenance management module 70 is used to perform pre-maintenance management of energy storage equipment according to the abnormal identification of the energy storage equipment. Specifically, the pre-maintenance management module 70 receives the abnormal identification of the energy storage equipment from the abnormal identification generation module 60 of the energy storage equipment, parses the received abnormal identification of the energy storage equipment, extracts key information, such as the number of the energy storage equipment, the type of abnormality, the time of occurrence, etc., analyzes this information, determines the nature and possible causes of the abnormality, and formulates a targeted pre-maintenance plan based on the analysis results of the abnormal information, combined with the historical maintenance records of the energy storage equipment, the type of equipment, the operating environment and other factors, including the time, location, personnel, required tools and materials, maintenance steps and other detailed information of the maintenance, organizes relevant personnel and equipment according to the requirements of the pre-maintenance plan, and performs preventive maintenance on the energy storage equipment, including inspection, replacement, repair and other operations. After the maintenance is completed, the maintenance process, results, materials used and time and other information are recorded, and the maintenance record of the energy storage equipment is updated. Through preventive maintenance, the occurrence rate and maintenance cost of energy storage equipment failures are reduced to ensure that the energy storage equipment can operate normally and extend its service life. The embodiment of the present application adopts the battery health status SOH value according to the energy storage device model, sorts the initial health factors by correlation, configures the timing prediction model based on the key health factors and the energy storage device model, builds the SOH value evaluation network, predicts the key health factors based on the number of charge and discharge cycles, generates the key health factor prediction status value, fits the key health factor prediction status value and the k-cycle label, generates the k-cycle SOH value and other technical means, achieves the technical effect of improving the efficiency and objectivity of the health factor evaluation and improving the accuracy of the battery health status evaluation results.

[0051] Although the present application makes various references to certain modules in the device according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0052] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A power management device for energy storage equipment, characterized in that: include: A battery service information acquisition module, the battery service information acquisition module is used to obtain battery service information, wherein the battery service information includes the number of charge and discharge cycles and the model of the energy storage device; A key health factor generation module, wherein the key health factor generation module is used to sort the initial health factors by correlation according to the battery health status SOH value of the energy storage device model to generate key health factors; An SOH value evaluation network construction module, wherein the SOH value evaluation network construction module is used to configure a time series prediction model based on the key health factors and the energy storage device model to construct an SOH value evaluation network; A key health factor prediction state value generation module, the key health factor prediction state value generation module is used to predict the key health factor based on the number of charge and discharge cycles, and generate a key health factor prediction state value, wherein the key health factor prediction state value has a k-cycle label; A k-cycle SOH value generation module, wherein the k-cycle SOH value generation module is used to fit the key health factor prediction state value and the k-cycle label through the SOH value evaluation network to generate a k-cycle SOH value; An energy storage device abnormal identification generation module, the energy storage device abnormal identification generation module is used to generate an energy storage device abnormal identification based on the k-cycle label when the k-cycle SOH value does not meet the k-cycle SOH threshold; A pre-maintenance management module, the pre-maintenance management module is used to perform pre-maintenance management of the energy storage device according to the abnormal identification of the energy storage device; Wherein, the key health factor generation module includes: A retrieval unit, the retrieval unit is used to retrieve the battery state of health SOH value and the initial health factor based on the energy storage device model as a constraint, and generate an SOH retrieval value set and an initial health factor retrieval value set, wherein the initial health factor retrieval value set corresponds to the SOH retrieval value set one by one; A dimension removal processing unit, the dimension removal processing unit is used to perform dimension removal processing on the SOH search value set, construct a reference SOH value sequence, traverse the initial health factor search value set to perform dimension removal processing, and construct multiple comparison health factor value sequences; A grey correlation evaluation unit, the grey correlation evaluation unit being used to perform grey correlation evaluation based on the reference SOH value sequence and the plurality of comparison health factor value sequences to generate a set of correlation evaluation values, wherein the set of correlation evaluation values ​​corresponds one-to-one to the initial health factors; A key health factor extraction unit, the key health factor extraction unit being used to extract the key health factors having a degree of association evaluation value greater than or equal to a degree of association evaluation threshold from the initial health factors based on the degree of association evaluation value set; The initial health factor construction unit includes at least the trickle current rise time, the constant current rise time, the constant voltage drop time, the cycle end voltage recovery value, the battery charging temperature, and the battery discharging temperature.

2. The power management device for energy storage equipment according to claim 1, characterized in that: The retrieval unit comprises: A retrieval subunit, the retrieval subunit is used to retrieve the battery state of health SOH value and the initial health factor based on the energy storage device model as a constraint, and obtain a first SOH retrieval value set and a first initial health factor retrieval value set; an aggregation subunit, the aggregation subunit being used to perform aggregation analysis on the first initial health factor retrieval value set to generate a plurality of clusters of first initial health factor retrieval values, wherein the plurality of clusters of first initial health factor retrieval values ​​have a centroid initial health factor retrieval value; A grouping subunit, the grouping subunit is used to group the first SOH search value set according to the multiple clusters of first initial health factor search values ​​to generate multiple clusters of first SOH search values; A de-extreme value fitting subunit, the de-extreme value fitting subunit is used to traverse the first SOH search values ​​of the multiple clusters to perform de-extreme value fitting and generate SOH fitting values; A retrieval value set generating subunit is used to set the centroid initial health factor retrieval value as the initial health factor retrieval value set, and set the SOH fitting value as the SOH retrieval value set.

3. The power management device for energy storage equipment according to claim 2, characterized in that: The polymerizing subunit comprises: an initial health factor retrieval value obtaining microunit, wherein the initial health factor retrieval value obtaining microunit is used to obtain a first group of initial health factor retrieval values ​​of the first initial health factor retrieval value set, until an Nth group of initial health factor retrieval values ​​of the first initial health factor retrieval value set is obtained; A normalization processing micro-unit, the normalization processing micro-unit is used to perform normalization processing on the first group of initial health factor retrieval values ​​to construct a first coordinate, until the Nth group of initial health factor retrieval values ​​is normalized to construct an Nth coordinate; A cluster analysis micro-unit, the cluster analysis micro-unit is used to perform cluster analysis on the first coordinate to the Nth coordinate based on a Euclidean distance threshold to generate a multi-cluster coordinate clustering result; A centroid sorting micro-unit, wherein the centroid sorting micro-unit is used to traverse the multi-cluster coordinate clustering results to perform centroid sorting and generate a plurality of centroid coordinates; A set grouping micro-unit, the set grouping micro-unit is used to group the first initial health factor retrieval value set according to the multi-cluster coordinate clustering result to generate the multi-cluster first initial health factor retrieval value; A set sorting micro unit is used to sort the first initial health factor retrieval value set according to the multiple centroid coordinates to generate the centroid initial health factor retrieval value.

4. The power management device for energy storage equipment according to claim 3, characterized in that: The centroid sorting microunit comprises: A coordinate acquisition sub-micro unit, wherein the coordinate acquisition sub-micro unit is used to obtain the first cluster intra-coordinates of the first cluster coordinate clustering result up to the Mth cluster intra-coordinates; A coordinate set acquisition sub-micro unit, the coordinate set acquisition sub-micro unit is used to acquire a first reference coordinate set with the coordinates in the first cluster as a reference point and a preset number of neighboring coordinates as a constraint, until an Mth neighboring coordinate set is acquired with the coordinates in the Mth cluster as a reference point and the preset number of neighboring coordinates as a constraint, to obtain a second reference coordinate set with the coordinates in the first cluster as neighboring coordinates; A centroid acquisition sub-micro unit, the centroid acquisition sub-micro unit is used to calculate the inverse of the mean of the distances from the first reference coordinate set and the second reference coordinate set to the first cluster internal coordinates, and set it as the first centroid, until the Mth centroid is obtained; The centroid coordinate generation sub-micro unit is used to set the coordinates corresponding to the maximum values ​​of the first centroid degree to the Mth centroid degree as the first cluster centroid coordinates and add them into the multiple centroid coordinates.

5. The power management device for energy storage equipment according to claim 1, characterized in that: The SOH value evaluation network building module includes: An online charging history curve sequence acquisition unit, the online charging history curve sequence acquisition unit is used to collect online charging history curve sequences based on the key health factors and the energy storage device model as constraints, wherein the charging history curve sequences correspond to the number of charge and discharge cycles one by one; A convolution feature extractor construction unit, the convolution feature extractor construction unit is used to construct a convolution feature extractor, wherein the convolution feature extractor is used to process the charging history curve sequence to extract the key health factor time series characteristics of the key health factor; A feature extraction unit, the feature extraction unit is used to extract features from the charging history curve sequence through the convolution feature extractor to obtain key health factor time series feature record data, wherein the number of charge and discharge cycles has a SOH true value label sequence; A splitting unit, the splitting unit is used to split the key health factor time series feature record data to obtain a first group of key health factor feature sequences, wherein the first group of key health factor feature sequences has a sequence end cycle number label, and the first group of key health factor feature sequences has at least three sequence nodes; An SOH true value label sequence sorting unit, the SOH true value label sequence sorting unit is used to sort the SOH true value label sequence according to the first group of key health factor feature sequences to obtain a sequence endpoint SOH true value label; A training data set construction unit, wherein the training data set construction unit is used to use the first group of key health factor feature sequences and the sequence endpoint cycle number label as inputs of the long short-term memory neural network, use the sequence endpoint SOH true value label as supervision of the long short-term memory neural network, construct a training data set, and configure the SOH value evaluation network.

6. The power management device for energy storage equipment according to claim 1, characterized in that: The key health factor prediction state value generation module includes: A local charging history curve sequence acquisition unit, the local charging history curve sequence acquisition unit is used to acquire a local charging history curve sequence according to the number of charge and discharge cycles; A key health factor fluctuation curve construction unit, wherein the key health factor fluctuation curve construction unit is used to construct a key health factor fluctuation curve according to the local charging history curve sequence; A segmentation unit, the segmentation unit is used to segment the key health factor fluctuation curve according to the charging mode to obtain a segmentation result of the key health factor fluctuation curve; A health factor average change speed extraction unit, the health factor average change speed extraction unit is used to traverse the key health factor fluctuation curve segmentation results and extract the health factor average change speed, wherein the health factor average change speed corresponds to the charging mode one by one; A cycle label configuration unit, the cycle label configuration unit is used to configure the k-times cycle label through a user terminal; A predicted state value generating unit, the predicted state value generating unit is used to predict the key health factor based on the k-cycle label and the average change rate of the first health factor in the charging mode with the maximum SOH value loss, and generate a first predicted state value; based on the k-cycle label and the average change rate of the second health factor in the charging mode with the minimum SOH value loss, the key health factor is predicted to generate a second predicted state value; A predicted state interval construction unit, the predicted state interval construction unit is used to construct a predicted state interval according to the first predicted state value and the second predicted state value, and add the predicted state interval to the key health factor predicted state value.