Online energy consumption forecast for battery electric buses using a learning-free algebraic method.

Area of Science:

• Transportation Engineering
• Energy Systems Analysis
• Computational Mathematics

Background:

• Accurate energy consumption prediction is crucial for efficient route planning and deployment of battery electric buses (BEBs).
• Current machine learning (ML)-based methods often require extensive historical data and computationally intensive offline training.
• The adaptability of prediction models to new driving conditions and transit modes remains a challenge.

Purpose of the Study:

• To develop and evaluate a novel, learning-free algebraic method for real-time energy consumption forecasting in BEBs.
• To demonstrate the method's independence from historical data and offline training requirements.
• To assess the proposed method's efficiency, adaptability, and performance compared to ML-based approaches.

Main Methods:

• The proposed method utilizes algebraic derivative estimation for real-time energy consumption prediction.
• It operates online, involving only algebraic calculations, thus eliminating the need for historical data and offline training.
• Performance is validated through comprehensive comparisons with representative ML-based methods using real-world data.

Main Results:

• The learning-free algebraic method achieves accurate real-time energy consumption predictions without historical data.
• It demonstrates superior calculation efficiency due to its online algebraic computation.
• The method exhibits strong adaptability to novel driving cycles and emerging transit services.

Conclusions:

• The proposed algebraic method offers a computationally efficient and data-independent alternative for BEB energy consumption prediction.
• Its adaptability makes it particularly suitable for dynamic transportation scenarios like on-demand transit.
• The study highlights the potential of algebraic methods to overcome limitations of traditional ML approaches in this domain.