By Anton Shvets, Jaichitra Balakrishnan, & Paul Tluczek

The BMW ChargeForward project in collaboration with California-based utility PG&E is designed to optimize electric vehicle (EV) charging.
It also explores the following topics:

  • Using vehicle telematics to measure performance data
  • Evaluating charging events to determine if they offer additional grid benefits beyond overnight charging
  • Exploring additional optimization opportunities using renewable energy sources

Our analysis is divided into following sections:

  • User Energy Consumption Analysis
  • Area Average Consumption Statistical Distribution
  • User Energy Consumption Segmentation
  • EV Charge Time Predictive Modeling
  • Vehicle Driving Behavior Modeling
  • Vehicle Plug-in Pattern Statistical Modeling
  • EV Charge and Grid Optimizer Model Using Renewable Energy Sources

Activities in the project include:

  • Data Scraping, Data Wrangling, Data Cleaning, Web Scraping
  • Statistical Data Analysis and Fitting Distributions
  • Geolocation Analysis
  • Pattern Identification Algorithm
  • Machine Learning Modeling – Classification
  • Machine Learning Modeling – Regression
  • Machine Learning Modeling – Clustering

Our analysis and research gives BMW insight for implementations of EV charging recommendation engines. The next phase of BMW ChargeForward applies our findings to advance their real-time vehicle technical information with predictions of travel behavior, grid load, and energy market price analysis. This strategically enables managing a vehicle’s charging through its current and projected charging events.