Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Battery Discharge Modeling for Electric Vehicles: A Hybrid Physics-based Residual Learning Approach

Praharshitha Aryasomayajula, Ting Bai, Andreas A. Malikopoulos

Abstract

The growing integration of electric vehicle (EV) fleets into transportation services and energy systems requires accurate modeling of battery discharge and state-of-charge (SoC) evolution to ensure reliable vehicle operation and grid coordination. Existing approaches face a trade-off between interpretable but simplified physics-based models and data-driven methods that demand large datasets and may lack physical consistency. In this paper, we propose a hybrid physics-based residual learning framework for EV battery discharge modeling. A vehicle dynamics model based on force-balance equations provides an interpretable baseline estimate of energy consumption and SoC evolution, capturing aerodynamic drag, rolling resistance, and regenerative braking. A neural network residual learner then corrects discrepancies caused by complex factors such as traffic conditions and driver behavior. Experimental results on $1,500$ trip scenarios demonstrate that the proposed approach reduces the mean absolute percentage error to approximately $0.8\%$, significantly outperforming physics-only models while preserving physical interpretability and computational efficiency.

Battery Discharge Modeling for Electric Vehicles: A Hybrid Physics-based Residual Learning Approach

Abstract

The growing integration of electric vehicle (EV) fleets into transportation services and energy systems requires accurate modeling of battery discharge and state-of-charge (SoC) evolution to ensure reliable vehicle operation and grid coordination. Existing approaches face a trade-off between interpretable but simplified physics-based models and data-driven methods that demand large datasets and may lack physical consistency. In this paper, we propose a hybrid physics-based residual learning framework for EV battery discharge modeling. A vehicle dynamics model based on force-balance equations provides an interpretable baseline estimate of energy consumption and SoC evolution, capturing aerodynamic drag, rolling resistance, and regenerative braking. A neural network residual learner then corrects discrepancies caused by complex factors such as traffic conditions and driver behavior. Experimental results on trip scenarios demonstrate that the proposed approach reduces the mean absolute percentage error to approximately , significantly outperforming physics-only models while preserving physical interpretability and computational efficiency.
Paper Structure (24 sections, 33 equations, 3 figures, 9 tables)

This paper contains 24 sections, 33 equations, 3 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Physics-Based Discharge Model
  3. Problem Formulation
  4. Driving Behavior Profiles
  5. Eco driving
  6. Normal driving
  7. Aggressive driving
  8. Power Consumption Model
  9. Computational Implementation
  10. Physics-based Model Validation
  11. Learning-Based Residual Correction Model
  12. Motivation and Approach
  13. Neural Network Architecture
  14. Training Data and Procedure
  15. Hybrid Physics-ML Approach
  16. ...and 9 more sections

Figures (3)

  • Figure 1: Comparison of SoC depletion between the physics-based model and baseline methods.
  • Figure 2: Prediction error distributions for physics-only, pure ML, and hybrid models (a--c), and per-style error breakdown for the hybrid model (d).
  • Figure 3: Energy consumption rate vs. trip distance.

Theorems & Definitions (1)

  • Remark 1