Sign in Subscribe
Machine learning

Machine Learning-Based Approach Reduces Energy Consumption in Electric Vehicles

Strategic Insights

— 2 min read

A new study has introduced a novel machine learning-based strategy for generating supercapacitor reference current to optimize energy distribution in Battery Electric Vehicles (BEV) and Hybrid Battery Electric Vehicles (HBEV). Researchers from the School of Electrical Engineering have developed a Long Short-Term Memory (LSTM) neural network trained using real-world drive cycle data, which is exported in Open Neural Network Exchange (ONNX) format for real-time deployment within a Simulink-based control environment.

Key Takeaways:

  • The LSTM-ONNX framework reduces battery stress, improves thermal performance, and enhances energy efficiency in electric vehicles by 21.3% in EUDC cycle, 18.1% in peak power demand, and 5.75% in battery energy consumption.
  • The proposed approach also leads to smoother traction motor torque, reduced current ripple, and optimized power delivery in hybrid energy storage electric vehicles.
  • The system is modeled using the Nissan Sakura EV and evaluated under EUDC and IM240 drive cycles, with significant improvements in battery current, power demand, and energy consumption.
  • The research introduces a novel machine learning-based strategy for generating supercapacitor reference current to optimize energy distribution in BEV and HBEV.
  • The LSTM neural network is trained using real-world drive cycle data and exported in ONNX format for real-time deployment within a Simulink-based control environment.
  • The research concludes that the proposed LSTM-ONNX framework is applicable for reducing battery stress, improving thermal performance, and enhancing energy efficiency in electric vehicles.
  • The School of Electrical Engineering has developed a comprehensive model to confirm the real-time applicability of the data-driven control strategy for electric vehicles.

Statistics:

  • 21.3% reduction in battery peak current in EUDC cycle
  • 18.1% reduction in peak power demand in EUDC cycle
  • 5.75% lower battery energy consumption in EUDC cycle
  • 33.5% reduction in peak battery current in IM240 cycle
  • 31.6% reduction in peak power in IM240 cycle
  • 12.36% reduction in battery energy consumption in IM240 cycle
  • 21.3% reduction in battery peak current (EUDC cycle)
  • 18.1% reduction in peak power demand (EUDC cycle)
  • 5.75% lower battery energy consumption (EUDC cycle)
  • 33.5% reduction in peak battery current (IM240 cycle)
  • 31.6% reduction in peak power (IM240 cycle)
  • 12.36% reduction in battery energy consumption (IM240 cycle)

Sources:

  • Machine learning-based approach for reduction of energy consumption in hybrid energy storage electric vehicle. Scientific Reports, 2025;15(1):29303.
  • Nature Portfolio, Heidelberger Platz 3, Berlin, 14197, Germany (www.nature.com/; www.nature.com/srep/).
  • T. Paulraj, School of Electrical Engineering, Vellore Institute of Technology, Vellore, Tamil Nadu, India.
  • NewsRx LLC, Copyright 2025.
  • Findings from School of Electrical Engineering Broadens Understanding of Machine Learning (Machine learning-based approach for reduction of energy consumption in hybrid energy storage electric vehicle). Journal of Engineering. August 25, 2025; p 690.

Read more