Hierarchical Power Allocation Method Based on Online Extremum Seeking Algorithm for Dual-PEMFC/Battery Hybrid Locomotive

This paper presents a hierarchical power allocation method (HPAM) for dual-stack fuel cell (FC) hybrid locomotive powertrain. In a high-power FC application such as rail transportation, the use of multi-stack system is able to provide the primary and stable demanded power. The battery is adopted as the energy storage source (ESS) to fulfill the transient power demand fluctuations. Besides, considering that the performance of the stack will change with the operating conditions, this study also searches an online identification method based on forgetting factor recursive least square (FFRLS) algorithm to update the system parameters of the dual-PEMFC system in real time. In addition, sequential quadratic programming (SQP) scheme is employed in this study to search for the optimal solution of nonlinear constraint problems. Furthermore, it is also to ensure that the proton exchange membrane fuel cell (PEMFC) systems can operate as much as possible in the high efficiency range, and the final battery state of charge (SOC) can be close to the initial SOC. Moreover, this study also adopts the equivalent consumption minimization strategy (ECMS) as benchmark to foreground the superiority of the presented method. A reduced-scale hybrid locomotive powertrain test bench is developed to demonstrate the practicability of the developed HPAM. It can be seen from the experimental results that, the proposed HPAM can optimize PEMFC systems efficiency and performances, and reduce system fuel consumption.

Automated summary

This paper introduces a hierarchical power allocation method for dual-stack fuel cell hybrid locomotive powertrain, utilizing multi-stack system and batteries as energy storage source. The study incorporates online identification and SQP scheme to update system parameters and find optimal solutions for efficient operation, along with using ECMS as a benchmark. Experimental results show that the proposed method can optimize system efficiency and reduce fuel consumption.