Predicting Life of Electric Vehicle and Load-Leveling Lead Acid Batteries from Initial Acceptance Test Data by Use of Pattern Recognition Analysis.

Abstract

A novel approach to battery lifetime prediction was applied to life-cycling data for 108 ESB EV-106 6-V golf cart batteries (tests conducted by TRW for NASA-Lewis). Computerized pattern recognition methods were used to examine initial cycling measurements and to classify individual batteries into long-lived or short-lived classes with greater than 85% accuracy. Results of this study were used to design a fabrication and test program for 340 GNB lead-acid cells in a 500 kWh load-leveling battery to be tested at the Battery Energy Storage Test Facility. Preliminary observations on initial test data are reported here. Keywords: Multivariate analysis.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1984
Accession Number
ADA168853

Entities

People

  • S. P. Perone
  • W. C. Spindler

Organizations

  • Lawrence Livermore National Laboratory

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Acceptance Tests
  • Acids
  • Data Mining
  • Databases
  • Electric Vehicles
  • Energy
  • Energy Storage
  • Engineering
  • Failure Mode And Effect Analysis
  • Information Science
  • Lead Acid Batteries
  • Materials
  • Materials Science
  • Military Research
  • Pattern Recognition
  • Production
  • Test Facilities

Readers

  • Battery Technology and Engineering
  • Regression Analysis.

Technology Areas

  • AI & ML