EP-4737927-A1 - SERVER AND OPERATION METHOD THEREOF

EP4737927A1EP 4737927 A1EP4737927 A1EP 4737927A1EP-4737927-A1

Abstract

According to an embodiment disclosed herein, a server may include a communication circuit, a memory, and a processor operatively connected to the communication circuit and the memory, wherein the processor may be configured to acquire pieces of time series data for a plurality of factors related to a battery for each user, generate feature data representing the relationship between at least two factors that are expected to have a mutual correlation among the plurality of factors based on each time series data, and classify the pieces of feature data into a plurality of clusters and define a plurality of patterns based on characteristics of each cluster.

Inventors

KIM, KANG SAN
KIM, GEUM BEE
YOO, JI WON
HEO, JUNG EUN

Assignees

LG Energy Solution, Ltd.

Dates

Publication Date: 20260506
Application Date: 20240812

Claims (14)

A server comprising: a communication circuit; a memory; and a processor operatively coupled to the communication circuit and the memory, wherein the processor is configured to: acquire pieces of time series data for a plurality of factors related to a battery for each user; generate feature data representing the relationship between at least two factors that are expected to have a mutual correlation among the plurality of factors based on each time series data; and classify the pieces of feature data into a plurality of clusters and define a plurality of patterns based on characteristics of each cluster.
The server of claim 1, wherein the processor generates the feature data based on a joint probability density function (joint PDF) of the at least two factors.
The server of claim 2, wherein the processor is configured to: calculate a joint probability mass function (joint PMF) for the at least two factors from the time series data; calculate the joint PDF by applying a kernel density function (KDE) to the joint PMF; and generate the feature data representing a joint probability density of the at least two factors.
The server of claim 3, wherein the feature data includes two-dimensional (2D) contour image data representing the joint probability density of the at least two factors.
The server of claim 1, wherein the processor classifies the pieces of feature data into the plurality of clusters based on the relationship between the at least two factors derived from the feature data and the distribution of each of the at least two factors.
The server of claim 1, wherein the processor is configured to: fuse the pieces of feature data included in each cluster for each cluster to generate representative feature data; and define the plurality of patterns based on the relationship between the at least two factors derived from the representative feature data and the distribution of each of the at least two factors.
The server of claim 1, wherein the processor determines a plurality of types for classifying a user based on the plurality of defined patterns.
The server of claim 7, wherein the processor analyzes the relationship with an evaluation index of the battery according to the plurality of types and extracts a main factor related to the evaluation index among the plurality of factors.
The server of claim 1, wherein the plurality of factors include at least some of a charging start state of charge (SOC), a charging depth of discharge (DOD), a charging C-rate, a charging start temperature, a highest charging temperature, a charging start time, a charging start day, a traveling start time, a traveling distance, a traveling start temperature, a highest traveling temperature, traveling DOD, traveling speed distribution, a discharging time, and a discharging end day.
The server of claim 1, wherein the processor acquires the time series data from an on board device (OBD) provided in each user's vehicle.
A method of operating a server, comprising: acquiring pieces of time series data for a plurality of factors related to a battery for each user; generating feature data representing the relationship between at least two factors that are expected to have a mutual correlation among the plurality of factors based on each time series data; and classifying the pieces of feature data into a plurality of clusters and defining a plurality of patterns based on characteristics of each cluster.
The method of claim 11, wherein the generating of the feature data includes: calculating a joint probability mass function (joint PMF) for the at least two factors from the time series data; calculating a joint probability density function (joint PDF) for the at least two factors by applying a kernel density function (KDE) to the joint PMF; and generating the feature data representing a joint probability density of the at least two factors.
The method of claim 11, wherein the defining of the plurality of patterns includes: classifying the pieces of feature data into the plurality of clusters based on the relationship between the at least two factors derived from the feature data and the distribution of each of the at least two factors; fusing the pieces of feature data included in each cluster for each cluster to generate representative feature data; and defining the plurality of patterns based on the relationship between the at least two factors derived from the representative feature data and the distribution of each of the at least two factors.
The method of claim 11, further comprising determining a plurality of types for classifying a user based on the plurality of defined patterns.

Description

[TECHNICAL FIELD] CROSS-REFERENCE TO RELATED APPLICATION This application claims priority to and the benefit of Korean Patent Application No. 10-2023-0114896 filed in the Korean Intellectual Property Office on August 30, 2023, the entire contents of which are incorporated herein by reference. TECHNICAL FIELD Embodiments disclosed herein relate to a server and a method of operating the same. [BACKGROUND ART] Recently, research and development on secondary batteries has been actively conducted. Here, the secondary battery is a battery capable of charging and discharging and includes all recent lithium ion batteries such as conventional Ni/Cd batteries and Ni/MH batteries. Among the secondary batteries, the lithium-ion batteries have an advantage of having a much higher energy density than the conventional Ni/Cd batteries, the Ni/MH batteries, and the like. In addition, since the small and light-weight lithium ion batteries may be manufactured, the lithium ion batteries are used as a power source for mobile devices, and recently, the lithium ion batteries are attracting attention as a next-generation energy storage medium due to the expansion of the range of use to a power source for electric vehicles. When such a battery is provided in a vehicle, it is common to collect and analyze battery data because there may be individual differences in the performance, lifetime, safety, etc. of the battery depending on the tendency of a vehicle driver. However, there is a problem that a large amount of data collected for each vehicle are present, and it is difficult to standardize the collected data because battery usage patterns of vehicle users are different. [DISCLOSURE] [TECHNICAL PROBLEM] Embodiments disclosed herein are directed to providing a server and a method of operating the same, which may standardize a large amount of data and define a user's type considering composite factors. Embodiments disclosed herein are also directed to providing a server and a method of operating the same, which may extract main factors related to an evaluation index of a battery. The technical objects of the embodiments disclosed herein are not limited to the above-described technical objects, and other objects that are not described will be able to be clearly understood by those skilled in the art from the following descriptions. [TECHNICAL SOLUTION] According to an embodiment disclosed herein, a server may include a communication circuit, a memory, and a processor operatively connected to the communication circuit and the memory, wherein the processor may be configured to acquire pieces of time series data for a plurality of factors related to a battery for each user, generate feature data representing the relationship between at least two factors that are expected to have a mutual correlation among the plurality of factors based on each time series data, and classify the pieces of feature data into a plurality of clusters and define a plurality of patterns based on characteristics of each cluster. According to an embodiment, the processor may generate the feature data based on a joint probability density function (joint PDF) of the at least two factors. According to an embodiment, the processor may be configured to calculate a joint probability mass function (joint PMF) for the at least two factors from the time series data, calculate the joint PDF by applying a kernel density function (KDE) to the joint PMF, and generate the feature data representing a joint probability density of the at least two factors. According to an embodiment, the feature data may include two-dimensional (2D) contour image data representing the joint probability density of the at least two factors. According to an embodiment, the processor may classify the pieces of feature data into the plurality of clusters based on the relationship between the at least two factors derived from the feature data and the distribution of each of the at least two factors. According to an embodiment, the processor may be configured to fuse the pieces of feature data included in each cluster for each cluster to generate representative feature data, and define the plurality of patterns based on the relationship between the at least two factors derived from the representative feature data and the distribution of each of the at least two factors. According to an embodiment, the processor may determine a plurality of types for classifying a user based on the plurality of defined patterns. According to an embodiment, the processor may analyze the relationship with an evaluation index of the battery according to the plurality of types and extract a main factor related to the evaluation index among the plurality of factors. According to an embodiment, the plurality of factors may include at least some of a charging start state of charge (SOC), a charging depth of discharge (DOD), a charging C-rate, a charging start temperature, a highest charging temperature, a charging start time,