Based on battery property research, we realize innovative stability of all batteries
through the IT platform actually communicating with batteries.

The technology of estimating a state of battery and diagnosing battery anormality
with the use of the data acquired in real time

Battery anomaly diagnosis technology

Measurement of the operation conditions where battery cells and modules have anormality

Quantification of state change in battery anormality

Notice of abnormal cells through the tracking of normal ageing process

  • · Battery parameter
    · User’s driving pattern
    · Estimation property and state

  • Property analysis based anormal diagnosis logic

  • Precise diagnosis technology of anormality in a level of pack, module, or cell

Battery state estimation technology

High-precise SOC calculation through the real-time state estimation of a running battery

Precise estimation technology of battery deterioration according to battery driving conditions

Calculation of remaining available capacity (SoAE), available power (SoP), and high accuracy

  • · Battery parameter
    · Battery usage patterns

  • Property analysis based state estimation model

  • Tens of data driven battery state and property estimation

Data (data-driven) and modeling based state diagnosis

    Cell Model based Property

  • < Input Data Pre-processing >

  • arrow_forward
  • < Model Parameter Estimation >

    Cycle Data based Property

  • < Properties that can be derived when charging and discharging >

  • < Properties that can be derived at rest >

    AI Model based Property

Unique batter diagnosis property

    IP Model & Error Distribution

Multi-IP based diagnosis with high accuracy

    A variety of data driven battery state estimation model

  • • Various methodologies, such as deep learning model, KNN clustering, K-means, and Bayesian inference, and track record

    • Optimized diagnosis property (IP) and application experience according to diagnosis objects and conditions

  • Betterwhy’s unique, precise,
    and diverse diagnosis IPs

  • A variety of machine learning models
    by application condition

  • Unsupervised learning that requires
    no cell-by-cell data in advance

  • Probability based diagnosis
    supporting meticulous classification

  • arrow_downward

Implementation of short-time & high-precise diagnosis by applying specified IP and diagnosis techniques differently depending on conditions