Optimization based method to develop representative driving cycle for real-world fuel consumption estimation

The lack of representative driving cycles is cited as one of main reasons for the increasing gap between vehicle test cycle and real-world fuel consumptions. Many past studies employed random and semi random methods for developing driving cycles, by which the driving cycles aligned with real world driving characteristics may not be obtained. Besides, most of the existing methodologies were proposed for relative long trajectories, and cannot handle short trajectories chopped for road segments. Therefore, a new Simulated Annealing (SA) based method is proposed, resulting in a speed-acceleration pattern better aligned with real-world driving characteristics. The speed-acceleration status transitions are directly derived from the sample snippets rather than idealized trip trajectories based on SA optimization. In a case study in Fujian Province, China, the SA-based method could stably converge to observed values as the number of iterations increases and it greatly reduces the error by up to 23% over traditional methods. Finally, the accuracy of fuel consumption estimation is improved by imposing restriction on the starting and ending speeds of driving cycles. The method could improve fuel consumption estimation and also provide a better understanding on regional driving pattern, it can be used as a valuable tool for supporting energy and climate policies. (c) 2021 Published by Elsevier Ltd.

Automated summary

The lack of representative driving cycles is a main reason for the increasing gap between vehicle test cycle and real-world fuel consumptions. A new Simulated Annealing (SA) based method is proposed to better align speed-acceleration pattern with real-world driving characteristics, which greatly reduces errors and improves fuel consumption estimation. It can serve as a valuable tool for supporting energy and climate policies.