maengjulie / battery-degradation-analysis-project-code

Battery analysis project

Analysis-of-the-Degradation-Patterns-of-a-Battery

w. Solvay Korea, Ewha Research Team

Data Deonising

tsmoothie_denoising.py: Denoising ("smoothing") the battery retention curves using a LOWESS smoother (Python tsmoothie package)

• Input: Raw battery data containing columns ‘battery_file_id’, 'DischargeCapacityRetention', 'Cyc_’, etc.
• Output: Plot visualizations of discharge capacity retention for all battery ids and the csv files of the smoothing results. The original curves are in blue, and the smoothed curves are red.
• Output example
  

Statistical Test

statistical_test.py: Performing Fischer’s exact test (or chi-squared test)

• Input: profiling data that was put together for descriptive analysis of formed clusters
• Output: Table images (values of significance highlighted in green)
• Output example

Oversampling

SMOTEN_oversampling.py: Oversampling the clustered data while accounting for component distributions (Python SMOTEN package)

• Input: Clustering result with meta data containing battery id, cluster, and components
• Output: 29 plots and 1 csv file
• Output example