Search

CN-121995260-A - Method and system for evaluating battery health state of power supply device of electric tool

CN121995260ACN 121995260 ACN121995260 ACN 121995260ACN-121995260-A

Abstract

The invention relates to the technical field of battery management and discloses a battery health state evaluation method and a system of an electric tool energy supply device, wherein the battery health state evaluation method of the electric tool energy supply device comprises the steps of acquiring multi-source sensor data in the running process of an electric tool and preprocessing; identifying and classifying multi-stage pulse discharge events, extracting multi-scale voltage response characteristics and performing temperature compensation, performing equivalent circuit model parameter identification, checking and eliminating abnormal identification results, extracting parameter aging characteristics, constructing a time sequence prediction model for prediction, evaluating the overall health state and single body consistency of a battery pack, defining a judgment standard of effective circulation, predicting the number of remaining effective circulation, calculating a confidence interval of a prediction result, generating a battery management suggestion, formatting into a standardized evaluation report, performing data synchronization and model optimization through edge cloud cooperation, and recording a system operation log.

Inventors

  • CHEN ZHENGGUAN

Assignees

  • 宁波微润精密机械有限公司

Dates

Publication Date
20260508
Application Date
20260116

Claims (10)

  1. 1. A method for evaluating the state of health of a battery of an energy supply device of an electric tool, comprising the steps of: acquiring multi-source sensor data in the running process of the electric tool and preprocessing the multi-source sensor data to obtain a normalized time sequence data set; Based on the normalized time sequence data set, identifying and classifying multi-stage pulse discharge events, extracting multi-scale voltage response characteristics and performing temperature compensation to obtain a health characteristic vector after temperature compensation; Based on the normalized time sequence data set and the health feature vector after temperature compensation, carrying out equivalent circuit model parameter identification, checking and eliminating abnormal identification results, and extracting parameter aging features; Fusing the health feature vector after temperature compensation with the parameter aging feature, constructing a time sequence prediction model for prediction, and obtaining the health state estimation result of each single battery; Based on the health state estimation results of all the single batteries, the overall health state and single consistency of the battery pack are estimated, and the overall health state estimation results of the battery pack are output; Defining a judging standard of effective circulation based on the state of health estimation result of each single battery and the whole state of health estimation result of the battery pack, predicting the number of remaining effective circulation, calculating a confidence interval of the predicted result, and generating a battery management suggestion; And formatting the residual effective cycle number prediction result and the battery management suggestion into a standardized evaluation report, carrying out data synchronization and time sequence prediction model optimization through edge cloud cooperation, and recording a system operation log.
  2. 2. The method of claim 1, wherein the step of acquiring and preprocessing multi-source sensor data during operation of the power tool comprises: acquiring terminal voltage, charge-discharge current and surface temperature of a battery pack of the electric tool in real time, and driving analog-to-digital converters of all channels by adopting a uniform clock source; Denoising the original sensor signal by adopting a median filtering algorithm; Performing missing value detection and interpolation processing, and judging the missing value when the data point is null or exceeds a physical reasonable range or the change rate between adjacent data points exceeds a preset maximum change rate threshold; filling the detected missing value by adopting an interpolation method based on local weighted regression; based on the load status identification signal, the continuous operational data is partitioned into independent duty cycle segments.
  3. 3. The method for evaluating the state of health of a battery of a power tool energy supply device according to claim 1, wherein the obtaining the temperature compensated health feature vector specifically comprises: Based on current time sequence data in the working period segment, identifying and classifying multi-stage pulse discharge events, calculating first-order and second-order differential sequences of current signals by adopting a multi-scale multi-condition joint judgment method, and searching positive jump points and negative jump points in the differential sequences; dividing the pulse into three grades of light-load pulse, medium-load pulse and heavy-load pulse according to the amplitude of the pulse current; respectively extracting multi-scale voltage response characteristics aiming at pulses of different grades, wherein the multi-scale voltage response characteristics comprise transient voltage sag amplitude, steady-state voltage sag amplitude, voltage sag rate, multi-stage voltage recovery time constant, hierarchical dynamic internal resistance and pulse energy efficiency; And calculating an equivalent value of the characteristic value at the reference temperature based on the battery temperature data and the temperature compensation model.
  4. 4. The method for evaluating the state of health of a battery of a power tool energy supply device according to claim 3, wherein the method for extracting the multi-stage voltage recovery time constant is as follows: And in the voltage recovery stage after the pulse is ended, fitting a recovery curve by adopting a double-exponential function model, carrying out parameter identification by adopting a nonlinear least square method, and adopting a Levenberg-Marquardt algorithm in the iteration process, wherein the convergence criterion is that the parameter variation of two adjacent iterations is smaller than a preset convergence threshold or the iteration times exceed a preset maximum iteration times.
  5. 5. The method for evaluating the state of health of a battery of a power tool supply device according to claim 1, wherein the extracting the parameter aging characteristic specifically comprises: establishing a second-order resistance-capacitance equivalent circuit model with self-adaptive pulse working conditions, introducing current dependence and frequency dependence correction, wherein the model structure comprises an ideal voltage source, current dependence ohmic internal resistance, a first resistance-capacitance parallel link and a second resistance-capacitance parallel link; based on a discretized state space model and measured voltage and current data, online identification of model parameters is carried out by adopting a recursive least square method, and a forgetting factor mechanism is introduced; carrying out validity check on the parameter estimation value to remove abnormal identification results, wherein the validity check adopts multiple judgment criteria, including physical range check, residual error check and change rate check; and extracting parameter aging characteristics based on the model parameter sequence subjected to the validity verification.
  6. 6. The method for evaluating the state of health of a battery of an energy supply device of an electric tool according to claim 1, wherein the obtaining the state of health evaluation result of each single battery specifically comprises: Carrying out feature fusion to construct model input, wherein the feature fusion adopts a vector splicing mode, and each feature is subjected to standardized treatment before splicing; Designing a structure of a long-period memory network, wherein a hidden layer of the network is formed by stacking long-period memory units, and each long-period memory unit comprises four core components including a forgetting door, an input door, an output door and a cell state; And (3) realizing uncertainty quantification by adopting an integrated prediction method, training a plurality of long-term and short-term memory network models with different initialization parameters, and estimating the prediction uncertainty by using the distribution of the prediction results of the plurality of models.
  7. 7. The method for evaluating the state of health of a battery of a power tool supply device according to claim 1, wherein the evaluation result of the overall state of health of the output battery pack specifically includes: Collecting voltage and temperature data of each single battery in the battery pack, and calculating a dispersion index among the single batteries; Analyzing the dispersion of the health state of the monomer, identifying an abnormal monomer, and judging the abnormal monomer by adopting a statistical threshold method when the deviation average value of the health state of a certain monomer exceeds a preset multiple threshold value; The method comprises the steps of constructing a graph neural network model to capture the mutual influence relation among monomers, modeling a battery pack as a graph, enabling each single battery to serve as a node of the graph, enabling the connection of sides to represent the association relation among the monomers, and enabling a forward propagation process of the graph neural network to adopt a message transmission mechanism; and designing a group-level health state aggregation function based on the updated feature vector of each monomer output by the graph neural network, and aggregating the health states of each monomer.
  8. 8. The method for evaluating the state of health of a battery of a power tool energy supply device of claim 1, wherein said generating a battery management advice specifically comprises: Defining a judging standard of the effective cycle, wherein the core index of the judging standard is the available energy of the battery, and judging the effective cycle when the available energy is not lower than the product of rated energy and an effective energy threshold coefficient; Predicting a time point when the health state falls to a failure threshold value by adopting an extrapolation method, selecting back-end data of a health state prediction curve as a fitting sample, and selecting an extrapolation model to fit the fitting sample; converting the failure time point into residual effective circulation times, counting historical use frequency of the battery, and converting the failure time point into residual circulation times according to the use frequency; comprehensively calculating a confidence interval of the residual effective cycle number prediction, wherein a Monte Carlo simulation method is adopted for the uncertainty comprehensive quantification; Based on the remaining valid cycle number prediction results and the confidence interval, a battery management recommendation is generated.
  9. 9. The method for evaluating the state of health of a battery of a power tool power supply device according to claim 1, wherein the formatting of the remaining effective cycle number prediction result and the battery management advice into a standardized evaluation report, the data synchronization and the time sequence prediction model optimization by the edge cloud cooperation, the recording of the system operation log specifically includes: formatting to generate a standardized evaluation report, wherein the evaluation report adopts a structured data format; Uploading data to a cloud platform through a wireless communication module, and performing aggregation storage and analysis processing after the cloud platform receives the uploaded data; Issuing model update to local equipment through a downlink channel, wherein the content of the model update comprises weight parameter update of a deep learning model, parameter adjustment of a feature extraction algorithm and an optimization value of a decision threshold; based on state information of each link in the evaluation process, a system operation log is recorded, and recorded contents of the log comprise a data acquisition state, a feature extraction result, a model reasoning state, a communication state and a system resource state.
  10. 10. A battery state of health evaluation system for a power tool power supply apparatus, for performing a battery state of health evaluation method of a power tool power supply apparatus as set forth in any one of claims 1 to 9, comprising: The data acquisition module is used for acquiring multi-source sensor data in the running process of the electric tool and preprocessing the multi-source sensor data to obtain a normalized time sequence data set; The characteristic extraction module is used for identifying and classifying multi-stage pulse discharge events based on the normalized time sequence data set, extracting multi-scale voltage response characteristics and performing temperature compensation to obtain a health characteristic vector after temperature compensation; the parameter identification module is used for carrying out equivalent circuit model parameter identification based on the normalized time sequence data set and the temperature compensated health feature vector, and verifying and eliminating an abnormal identification result to obtain a parameter aging characteristic; The health prediction module is used for fusing the health feature vector after temperature compensation with the parameter aging feature, constructing a time sequence prediction model for prediction, and obtaining the health state estimation result of each single battery; The consistency analysis module is used for evaluating the overall health state of the battery pack and consistency of the single batteries based on the health state estimation results of the single batteries and outputting the overall health state estimation results of the battery pack; The service life prediction module is used for defining a judgment standard of effective circulation based on the health state estimation result of each single battery and the overall health state estimation result of the battery pack, predicting the number of remaining effective circulation, calculating a confidence interval of the prediction result and generating a battery management suggestion; and the result output module is used for formatting the residual effective cycle number prediction result and the battery management suggestion into a standardized evaluation report, carrying out data synchronization and time sequence prediction model optimization through edge cloud cooperation, and recording a system operation log.

Description

Method and system for evaluating battery health state of power supply device of electric tool Technical Field The invention relates to the technical field of battery management, in particular to a battery health state evaluation method and system of an energy supply device of an electric tool. Background With the wide application of electric tools in the fields of industrial manufacture, building construction, equipment maintenance and the like, the importance of lithium ion batteries as core energy supply devices thereof is increasingly highlighted. The working characteristics of the electric tool determine that the battery pack of the electric tool faces multiple challenges such as high-frequency pulse discharge, high-current instantaneous load, intermittent working mode, complex and changeable environmental temperature conditions and the like, and the factors act together to enable the aging mechanism of the battery of the electric tool to present different characteristics from other application scenes such as an electric automobile, an energy storage system and the like, so that accurate evaluation of the health state of the battery has important significance for guaranteeing safe operation of equipment and optimizing battery asset management. Most of the existing battery state-of-health evaluation methods are developed based on standardized charge and discharge test conditions, and the batteries are assumed to work in a constant-current and constant-voltage mode, so that the method is difficult to adapt to nonstandard pulse working conditions in actual use of the electric tool. Meanwhile, the conventional method generally adopts general health indexes, and cannot fully mine aging characteristic information of the electric tool under special working conditions, so that evaluation accuracy is limited. In addition, the battery pack of the electric tool is composed of a plurality of single batteries in series-parallel connection, inconsistency among the single batteries is a key factor influencing group-level performance and service life, and the existing method lacks an effective mapping mechanism from single battery health status to group-level health status. Therefore, a battery health state evaluation method which can adapt to special working conditions of an electric tool, fully utilizes pulse discharge characteristic information and realizes cooperative evaluation from a single unit to a group level is needed, so as to solve the defects of the prior art in the aspects of working condition adaptability, characteristic extraction pertinence and single unit group level relevance. Disclosure of Invention The invention provides a battery health state evaluation method and system of an electric tool energy supply device, which solve the technical problems of insufficient working condition adaptability, feature extraction pertinence and monomer group level relevance in the related technology. The invention provides a battery health state evaluation method of an electric tool energy supply device, which comprises the following steps: acquiring multi-source sensor data in the running process of the electric tool and preprocessing the multi-source sensor data to obtain a normalized time sequence data set; Based on the normalized time sequence data set, identifying and classifying multi-stage pulse discharge events, extracting multi-scale voltage response characteristics and performing temperature compensation to obtain a health characteristic vector after temperature compensation; Based on the normalized time sequence data set and the health feature vector after temperature compensation, carrying out equivalent circuit model parameter identification, checking and eliminating abnormal identification results, and extracting parameter aging features; Fusing the health feature vector after temperature compensation with the parameter aging feature, constructing a time sequence prediction model for prediction, and obtaining the health state estimation result of each single battery; Based on the health state estimation results of all the single batteries, the overall health state and single consistency of the battery pack are estimated, and the overall health state estimation results of the battery pack are output; Defining a judging standard of effective circulation based on the state of health estimation result of each single battery and the whole state of health estimation result of the battery pack, predicting the number of remaining effective circulation, calculating a confidence interval of the predicted result, and generating a battery management suggestion; And formatting the residual effective cycle number prediction result and the battery management suggestion into a standardized evaluation report, carrying out data synchronization and time sequence prediction model optimization through edge cloud cooperation, and recording a system operation log. In a preferred embodiment, the acquiring and preprocessing m