Energy consumption characteristics based driving conditions construction and prediction for hybrid electric buses energy management. (15th April 2022)

• Record Type: Journal Article
• Authors: Wang, Yue; Li, Keqiang; Zeng, Xiaohua; Gao, Bolin; Hong, Jichao
• Is Part Of: Energy. Volume 245(2022)
• Publication Date: 2022-04-15
• Subjects: Hybrid electric buses -- Energy management -- Energy consumption characteristics -- Driving condition information; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Energy consumption -- Periodicals; 333.7905
• DOI: 10.1016/j.energy.2022.123189
• Languages: English
• ISSNs: 0360-5442

Abstract

Abstract: The energy consumption characteristics of driving condition is very important for hybrid electric buses energy management. In this paper, an energy consumption characteristics based driving conditions construction and prediction method was proposed. Under connected vehicular-cloud environment, missing data and noise data was processed by BP neural network method and wavelet transform method, respectively. According to the proposed the analysis method of energy consumption characteristics, the 7 characteristic parameters of the driving conditions related to energy consumption characteristics were extracted from 30 parameters. Based on the extracted characteristic parameters considering energy consumption, driving conditions construction and prediction were developed. In the driving cycle construction, it is found that the characteristic parameters error is less than 5% by comparing the original constructed cycles. Thus, the construction driving cycle can reflect the actual driving characteristics. In the driving cycle prediction, the prediction combining least squares support vector machine with BP neural network is proposed and compared with different topologies. The root mean square error of the proposed prediction model is 0.22 km/h, achieving the best prediction performance. Finally, energy management using the above driving condition information can significantly improve the energy economy. Highlights: Analysis of driving condition parameters considering energy consumption. Energy consumption characteristics based driving cycle construction and prediction. LS-SVM and BP-NN based intelligent prediction model of driving conditions. Missing and noise data were processed by BP-NN method and wavelet transform.