Search

Aiming at the problems that existing energy management strategies (EMS) are rarely applied to 100-ton loaders and the engine start-stops frequently under complex driving conditions, this paper proposes a novel EMS for 100-ton extended range hybrid loaders based on an artificial tree algorithm (AT). Firstly, using the equivalent fuel consumption minimization strategy (ECMS) as a foundation, a penalty function is designed to restrict the range extender's start-stop frequency and integrated into the ECMS control framework. Secondly, a real-time driving condition recognition model based on AT optimized back propagation (BP) Neural Network is proposed. Finally, with the equivalent factor, scale factor of state of charge (SOC) penalty function and range extender start-stop penalty function as optimization variables, and fuel economy, SOC stability, and range extender start-stop frequency as optimization objectives, AT is used for multi-objective optimization to obtain the optimal control parameters corresponding to the identified driving conditions. The simulation results demonstrate that compared with ECMS and Proportional-Integral-Derivative (PID) based ECMS, the proposed strategy is more effective for maintaining SOC stability. Besides, it improves the fuel economy by 5.937% and 1.353%, respectively, and decreases the number of range extender start-stops by 50.000% and 55.556%, respectively. • A novel energy management strategy applied to extended range hybrid loaders. • Integration of driving condition recognition and APU start-stop optimization. • A novel optimization strategy based on an artificial tree algorithm. • BP neural network optimized by artificial tree algorithm has better performance. • Fuel economy, SOC stability, and APU start-stop frequency have been improved. [ABSTRACT FROM AUTHOR]

The recent advancements in artificial intelligence have promoted deep reinforcement learning (DRL) as the preferred method for developing energy management strategies (EMSs) for fuel cell vehicles (FCVs). However, the development of DRL-based EMSs is a time-consuming process, requiring repetitive training when encountering different vehicle types or learning tasks. To surmount this technical barrier, this paper develops a transferable EMS rooted in heterogeneous deep transfer reinforcement learning (DTRL) across both FCV types and optimization tasks. Firstly, a simple source EMS based on the soft actor-critic (SAC) algorithm is pre-trained for a fuel cell sedan, solely focusing on hydrogen saving. After that, a heterogeneous DTRL framework is developed by integrating SAC with transfer learning, through which both heterogeneous deep neural networks and experience replay buffers can be transferred. Subsequently, the source EMS is transferred to the target new EMS of a fuel cell bus (FCB) to be reused, with additional consideration of the fuel cell (FC) longevity. Experimental simulations reveal that the heterogeneous DTRL framework expedites the development of the new EMS for FCB by 90.28 %. Moreover, the new EMS achieves a 7.93 % reduction in hydrogen consumption and suppresses FC degradation by 63.21 %. By correlating different energy management tasks of FCVs, this article both expedites the development and facilitates the generalized application of DRL-based EMSs. • A heterogeneous DTRL framework is designed by integrating TL with SAC algorithm. • A transferable EMS across FCV types and learning tasks is proposed based on DTRL. • Both heterogeneous DNNs and experience replay buffer are transferred and reused. • Fuel economy performance and fuel cell longevity are collaboratively optimized. • The proposed EMS shows great potential for real-time application via online test. [ABSTRACT FROM AUTHOR]

The propulsion and power systems are critical to the performance of unmanned aerial vehicles (UAVs). The rapid development of electrified propulsion systems powered by proton exchange membrane fuel cells (PEMFCs) has driven technological innovation of propulsion and power systems for UAVs. In this work, a 1.5 kW PEMFC integrated propulsion system with direct airflow intake design (Beihang Hydrogen-1) is developed and verified by flight test on a UAV with a wingspan of 3.6 m. The propulsion system can directly use the airflow behind the propeller to feed the PEMFC cathode for cooling and reaction and the power generated by the PEMFC is used to drive the propeller. A DC-DC module and a lithium battery pack are also added to the propulsion system to stabilize the output voltage and increase the instantaneous power for takeoff, transition, and landing stages. The "dynamic distribution" energy management strategy for the propulsion system is proposed to increase the system dynamic response ability. The detailed data of the PEMFC integrated propulsion system during the flight are measured and analyzed. The flight test continued for 18 min, with an average cruising altitude and an average cruising speed of 101 m above sea level and 23 m/s, respectively. During the cruise stage, the average PEMFC power is 1688 W, with an average single-cell voltage of 0.65 V. The average charging power of PEMFC for the lithium battery is 41 W. The successful flight test verifies the feasibility of the design of the PEMFC integrated propulsion system, provides valuable data for the following optimal design and flight test of the PEMFC integrated propulsion systems, and opens up a new direction for the development and application of PEMFC-powered electrified propulsion systems. • A 1.5 kW PEMFC integrated propulsion system with direct airflow intake design is developed. • The "dynamic distribution" energy management strategy is proposed for the PEMFC integrated propulsion system. • The first flight verification of the PEMFC integrated propulsion system is conducted based on a UAV. • The average PEMFC power is 1688 W with an average charging power of 41 W for the battery in the cruise stage of the flight test. • The PEMFC integrated propulsion system shows excellent dynamic response ability and energy conversion efficiency. [ABSTRACT FROM AUTHOR]

Generalization to unseen environments is still a challenge for deep reinforcement learning (DRL)-based energy management strategies (EMSs). This paper proposes a real-time EMS with high generalization for a light-duty hybrid electric vehicle (HEV) from two perspectives: enhancing the generalization of the DRL algorithm and improving the accuracy of application scenario representation in the training environment. The enhanced DRL algorithm named ATSAC can adjust the update frequency and learning rate of SAC automatically to improve the generalization. With the advancement of naturalistic driving big data (NDBD) and machine learning, a specific training cycle is synthesized based on NDBD to reflect an urban-suburban real-world driving scenario more accurately. By the comprehensive comparison with SAC and TD3 based EMSs applied to unseen driving scenarios, the proposed algorithm achieves significant improvement in computational efficiency, optimality, and generalization. The results show that the computational efficiency of ATSAC is increased by 52.32% compared to SAC. The negative total reward (NTR) of ATSAC is decreased by 18.22% and 69.81% compared to SAC and TD3, respectively. Further analysis shows that the EMS trained through the synthetic driving cycle obtains 18.37% lower NTR than WLTC which demonstrates that the synthetic method can reflect the state transition probability of real-world driving scenarios better than WLTC. • A novel DRL algorithm with high generalization is researched for energy management. • An SODC is synthesized through big data and machine learning as the training cycle. • Iteration dropout can avoid overfitting to improve the generalization of DRL. • Adaptive learning rate can balance the exploration and exploitation. [ABSTRACT FROM AUTHOR]