Measurement-based assessment and optimization of electric bus energy consumption under thermal variability using long-term field data
Measurement, cilt.280, sa.121795, ss.1-16, 2026 (SCI-Expanded)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 280 Sayı: 121795
- Basım Tarihi: 2026
- Doi Numarası: 10.1016/j.measurement.2026.121795
- Dergi Adı: Measurement
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED)
- Sayfa Sayıları: ss.1-16
- İnönü Üniversitesi Adresli: Evet
Özet
This study investigates the effect of ambient temperature on the energy consumption of overhead battery-hybrid electric buses, also known as trolleybuses. The analysis uses real field data. The data were collected from 22 hybrid buses over 24 months. Each bus was monitored during daily passenger service. The onboard black-box units recorded the main operating signals at 1 Hz. This allowed second-by-second changes in speed, braking, passenger load, and energy use to be examined under real route conditions. The study is not based on standard driving cycles or simulation data. It uses measured data from daily bus operation. The dataset includes energy consumption, regenerative braking, vehicle speed, passenger load, road gradient, auxiliary power demand, and ambient temperature. Before the analysis, missing, corrupted, and inconsistent records were removed. This step was used to improve the reliability of the measurement dataset. In this study, outside temperature is not treated as a direct electric motor parameter. It is considered a factor that affects total energy use through battery behavior and heating, ventilation, and air conditioning demand. Different prediction methods were tested on the processed data. These methods include Decision Trees, Ensemble Learning, Gaussian Process Regression, Support Vector Machines, and Three-Layer Neural Networks. The Three-Layer Neural Network optimized with the Modified Tunicate Swarm Algorithm achieved the best predictive performance. According to the analysis results, the model aimed to provide a practical basis for seasonal energy forecasting and route-based fleet planning by evaluating the sensitivity of energy consumption over a wide temperature range.