yjpark0421 / lstm-analytics-nasa-battery-ltv-prediction

PCoE Li-ion Battery Data Modeling

Remaining Useful Life Prediction Using NASA Battery Data Set

1️⃣ Overview

◽ Objective

• NASA에서 제공하는 4개 셀 리튬이온 배터리 데이터를 명세하고 분석하여 데이터를 이해한다.
• 데이터 전처리 및 학습에 알맞은 형태로 가공한다.
• AI(ML/DL)기반의 다양한 방법론(알고리즘,모델)을 적용한 배터리 잔존수명예측 모델링을 진행한다.

---

◽ Development Environment

page_type	languages	products
Dev	pyspark　python	azure　azure-databricks

---

◽ NASA PCoE Battery Data

Download URL - https://ti.arc.nasa.gov/tech/dash/groups/pcoe/prognostic-data-repository/

---

2️⃣ PreRequirement

◽ Understanding Data

• SOH : State of Health
• RUL : Remaining Useful Life
• EOL : End of Life

◽ Data Structure

• ❶ cycle: top level structure array containing the charge, discharge and impedance operations

  cycle	⇦ Complex Data Structure( Nested Data )

  ⏬

• ❷ type: operation type, can be charge, discharge or impedance

• ❸ Data: Nested Data Structure(Many Variables)

  type	ambient_temperature	time	data
  charge	24℃	yyyy-MM-dd HH:mm:ss.SSSSSS	data
  discharge	''	''	''
  impedance	''	''	''

---

3️⃣ Methodology

◽ Part1 : Convert .mat to Dataframe

• Convert .mat to Dataframe.ipynb

◽ Part2 : EDA(Exploratory Data Analysis)

• EDA-Battery-Data.ipynb

◽ Part3 : Modeling

• LSTM_Based_SOH_Prediction.ipynb
• SVM_Based_SOH_Prediction.ipynb

◽ LSTM Based Model Prediction Result Example

◽ Understanding LSTM(Long Short-Term Memory)

Read a Blog - http://colah.github.io/posts/2015-08-Understanding-LSTMs/

◽ Understanding SVM(Support Vector Machines)

Watch a Youtube - https://www.youtube.com/watch?v=eHsErlPJWUU

---

4️⃣ Reference

URL① - https://ieeexplore.ieee.org/document/8967059

lstm-analytics-nasa-battery-ltv-prediction's Issues

LSTM_Graph

NASA Battery Data

Long Short-Term Memory

python

Could you please provide the version of Python package？Thanks