CN-117783882-B - Battery pack temperature estimation method based on measurable parameter difference information

Abstract

A battery pack temperature estimation method based on measurable parameter difference information relates to the technical field of lithium ion batteries. Selecting a reference battery to calculate Euclidean distance between the terminal voltage of the single battery and all other single batteries provided with temperature sensors, determining a target battery, constructing input features and collecting, repeatedly selecting the reference battery to obtain a modeling data set, establishing a temperature estimation data driving model, calculating Euclidean distance between the terminal voltage of the single battery without the temperature sensors and all the single batteries provided with the temperature sensors to obtain an index value, calculating the input features and collecting based on the index value, and inputting the collection into the temperature estimation data driving model, wherein the measured temperature of the target battery is taken as an estimated temperature. And constructing input features based on the measurable parameter difference information, obtaining a modeling data set of the target battery index based on the distance index, further establishing a temperature estimation data driving model, improving robustness, and rapidly predicting a temperature estimation result through the target battery index value.

Inventors

• ZHAO LINHUI
• Qin Pengliang

Assignees

• 哈尔滨工业大学

Dates

Publication Date

20260512

Application Date

20231229

Claims (1)

1. 1. A battery pack temperature estimation method based on measurable parameter difference information is characterized by comprising the following steps: Step one, establishing a temperature estimation data driving model S1.1 modeling target acquisition for temperature estimation data driven model establishment Aiming at a battery pack with a temperature sensor arranged at a key position, selecting a single battery with the temperature sensor as a reference battery in the battery pack, and calculating the Euclidean distance of terminal voltage between the reference battery and all other single batteries with the temperature sensor, wherein the formula is as follows: where ED represents the calculation of the Euclidean distance, Represents the terminal voltage of the unit cell provided with the temperature sensor in addition to the reference cell, The terminal voltage of the reference battery is represented, and m represents the number of single batteries provided with a temperature sensor; and obtaining an index value index tar with the minimum distance difference through the calculated distance, wherein the formula is as follows: the single battery which is determined according to the index value index tar and is provided with the temperature sensor is used as a target battery for establishing a temperature estimation data driving model; s1.2 modeling feature construction for temperature estimation data driven model establishment The input feature 1 and the input feature 2 are respectively represented by the average value and the standard deviation of the sequence of the difference in heat generation amount between the unit cells, and are represented as follows: Fea 1 ＝mean(Q p -Q b ) (3) Fea 2 ＝std(Q p -Q b ) (4) Wherein Q p is the heat generation amount of the target battery, and Q b is the heat generation amount of the reference battery; the input features 3 and 4 are respectively represented by the average value and standard deviation of the series of terminal voltage differences between the unit cells, and are represented as follows: Fea 3 ＝mean(U t,p -U t,b ) (5) Fea 4 ＝std(U t,p -U t,b ) (6) The configuration of the input features 5 and 6 using the ratio between the terminal voltages is represented as follows: Fea 5 ＝mean(U t,p /U t,b ) (7) Fea 6 ＝std(U t,p /U t,b ) (8) wherein U t,p is the terminal voltage of the target battery, and U t,b is the terminal voltage of the reference battery; Taking the constructed input feature set as Fea= [ Fea 1 ,Fea 2 ,Fea 3 ,Fea 4 ,Fea 5 ,Fea 6 ] as a modeling feature for temperature estimation data driving model establishment; S1.3 modeling dataset construction for temperature estimation data driven model establishment Sequentially selecting the rest single batteries provided with temperature sensors in the battery pack as reference batteries, repeating the steps S1.1 and S1.2, and obtaining all modeling targets and modeling characteristics for temperature estimation data driving model establishment in the battery pack, wherein the formulas are as follows: wherein D mod is a modeling dataset; s1.4, establishing a temperature estimation data driving model Based on the obtained modeling data set, a temperature estimation data driving model is established by using a data driving algorithm, and the formula is as follows: M data ＝f(D mod ) (10) Wherein M data is a temperature estimation data driving model, and f is a data driving algorithm; step two, obtaining index values of single batteries without temperature sensors S2.1, obtaining and calculating Euclidean distances of terminal voltage between the single battery without the temperature sensor and all the single batteries with the temperature sensors in the battery pack by using index values of the single batteries with the minimum distance differences and the single batteries with the temperature sensors, wherein the Euclidean distances are expressed as follows: In the formula, The terminal voltage of the unit cell without the temperature sensor, The degree end voltage of the single battery provided with the temperature sensor is s, and the number of the single batteries without the temperature sensor is s; and obtaining an index value index tar,n with the minimum distance difference through the calculated distance, wherein the formula is as follows: The index value with the minimum distance difference between the single battery without the temperature sensor and the single battery with the temperature sensor is obtained, so that a basis is provided for temperature estimation of the subsequent single battery without the temperature sensor; S2.2, input feature calculation for temperature estimation based on measurable parameter difference information Based on the index value of the single battery with the minimum distance difference and provided with the temperature sensor obtained in the step S2.1, the input characteristics between the target battery corresponding to the obtained index value and the single battery without the temperature sensor are calculated as follows: Fea 1 '＝mean(Q n -Q p ) (13) Fea' 2 ＝std(Q n -Q p ) (14) Fea' 3 ＝mean(U t,n -U t,p ) (15) Fea' 4 ＝std(U t,n -U t,p ) (16) Fea' 5 ＝mean(U t,n /U t,p ) (17) Fea' 6 ＝std(U t,n /U t,p ) (18) Wherein Q n is the heat generation amount of the single battery without a temperature sensor, and U t,n is the terminal voltage of the single battery without a temperature sensor; Taking the calculated input characteristic set as Fea' = [ Fea 1 ',Fea' 2 ,Fea' 3 ,Fea' 4 ,Fea' 5 ,Fea' 6 ] as an input characteristic of single battery temperature estimation without a temperature sensor; Step three, estimating the temperature of the single battery without a temperature sensor Inputting the set Fea' calculated in the step S2.2 into the temperature estimation data driving model established in the step S1.4 to obtain a target battery index predicted value for estimating the temperature of the single battery without the temperature sensor, wherein the formula is as follows: index pre,n ＝M data (Fea') (19) in the temperature estimation, index pre,n corresponds to the measured temperature of the target battery, that is, the estimated temperature of the unit battery where the temperature sensor is not provided.

Description

Battery pack temperature estimation method based on measurable parameter difference information Technical Field The invention relates to the technical field of lithium ion batteries, in particular to a battery pack temperature estimation method based on measurable parameter difference information. Background The accurate estimation and effective monitoring of the temperature of each single battery in the battery pack have important significance for ensuring the thermal safety of the battery, preventing the occurrence of thermal runaway, safely driving the electric automobile and the like. However, most of the existing temperature estimation methods of the battery are proposed based on the single batteries, but the single batteries in the battery pack are numerous, the positions of the single batteries have large differences, and in addition, the influence of the cooling system on the temperature estimation of the battery is very remarkable. Most importantly, due to cost, space, and wiring, only a limited number of temperature sensors are placed on the cell surface with critical positions in the battery pack, while the temperature of other cells cannot be effectively monitored, and therefore, a method for effectively estimating the temperature of all cells in the battery pack is needed. Compared with the model-based method which mainly comprises a model-based method and a data-driven method, the model-based method is more widely favored in terms of development of technologies such as artificial intelligence, big data, internet of vehicles and the like, because the complicated heat generation and heat transfer mechanism of the battery, the position of the battery, the structure of a battery pack, whether a cooling system is started or not and the like are considered to establish a complicated temperature estimation model, and the data-driven method only needs to establish a mapping relation between parameters such as voltage, current, state of charge and the like and temperature. However, the data driving model established by the conventional data driving method depends on a large amount of modeling data, and although the temperature estimation result of the battery pack can be further improved based on limited measurable temperature information, the robustness of the established data driving model is poor due to inconsistent heat generation and heat transfer mechanisms of the single batteries at different positions in the battery pack. Disclosure of Invention In order to solve the defects in the background art, the invention provides a battery pack temperature estimation method based on measurable parameter difference information, which constructs input characteristics based on the measurable parameter difference information, obtains a modeling data set of a target battery index based on a distance index, further establishes a temperature estimation data driving model, is beneficial to improving robustness, and rapidly predicts a temperature estimation result through a target battery index value of a single battery provided with a temperature sensor. In order to achieve the purpose, the invention adopts the following technical scheme that the battery pack temperature estimation method based on the measurable parameter difference information comprises the following steps: Step one, establishing a temperature estimation data driving model S1.1 modeling target acquisition for temperature estimation data driven model establishment Aiming at a battery pack with a temperature sensor arranged at a key position, selecting a single battery with the temperature sensor as a reference battery in the battery pack, and calculating the Euclidean distance of terminal voltage between the reference battery and all other single batteries with the temperature sensor, wherein the formula is as follows: where ED represents the calculation of the Euclidean distance, Represents the terminal voltage of the unit cell provided with the temperature sensor in addition to the reference cell,The terminal voltage of the reference battery is represented, and m represents the number of single batteries provided with a temperature sensor; and obtaining an index value index tar with the minimum distance difference through the calculated distance, wherein the formula is as follows: the single battery which is determined according to the index value index tar and is provided with the temperature sensor is used as a target battery for establishing a temperature estimation data driving model; s1.2 modeling feature construction for temperature estimation data driven model establishment The input feature 1 and the input feature 2 are respectively represented by the average value and the standard deviation of the sequence of the difference in heat generation amount between the unit cells, and are represented as follows: Fea1＝mean(Qp-Qb) (3) Fea2＝std(Qp-Qb) (4) Wherein Q p is the heat generation amount of the target battery, and Q b is the heat genera