Search

• An online EMS based on random forest is proposed for fuel cell vehicles. • A LVQ classifier is designed to determine the driving pattern in real-time. • A NSGA-II based optimization method is used to optimize the EMS. • The method can reduce fuel cell degradation and exhibit good fuel economy. An effective energy management strategy (EMS) is crucial for ensuring the safe and efficient operation of fuel cell vehicles. This paper proposes an online EMS for the fuel cell/battery hybrid system based on an improved random forest (RF) algorithm. First, in order to enhance the environmental adaptability of EMS, a learning vector quantization classifier is utilized to identify the driving mode under different road conditions. In each driving mode, the optimal control sequence is derived using the Pontryagin's minimum principle (PMP), forming a globally optimal training set based on the idea of minimizing fuel consumption. Then an online application solution based on the RF is proposed to learn the optimal control sequence and address the real-time limitation. Furthermore, A RF hyper-parameter optimization method based on the NSGA-II algorithm is proposed to improve the accuracy and robustness of the RF under different driving patterns. The simulation experiments under two test conditions showed that compared with PMP method, the proposed EMS reached 97.67% and 98.71% of benchmark in terms of hydrogen consumption, and fuel cell degree of aging was reduced by 5.9% and 4.9%. Hence, the proposed method is proved to be capable of reducing degradation and exhibiting good fuel economy. [ABSTRACT FROM AUTHOR]

• Fast Tri-level optimization approach. • HRES includes PV panels, wind turbines and HESS. • Real weather and load data of the stand-alone HRES. • The optimal solution for sizing and EMS stages. • Minimize the capital cost and maximize the efficiency of the whole system. Renewable energy sources penetrate our recent life very deeply. They have more expansion chances in the nearest future, especially in stand-alone and micro grid fields. Due to the uncertainty of the weather and the non-matching between the generated energy and demand, a hybrid energy storage system (HESS) is required. The optimization computations are the best way to select the optimal values for the capacity, sizes and quantities of the stand-alone hybrid renewable energy system (HRES) to satisfy the load demands of the available climate conditions. In this paper, the HRES which includes PV panels, wind turbines, battery banks, hydrogen tanks with fuel cells and electrolyzers has been tested and optimized using a tri-level optimization problem. Using Matlab software, the simulation results validate that the proposed method converged quickly and accurately to the optimal solution. The total capital cost was between $127,400 to $231,750. The maximum average efficiency was 83.53% at a minimum total capital cost of $127,400 and a minimum operating cost of $67 which were achieved in the same summer month presented in July. [ABSTRACT FROM AUTHOR]

• Deep stochastic reinforcement learning based approach to address this issue of epistemic uncertainty in a midsize fuel cell hybrid electric vehicle. • The performance of the proposed approach is benchmarked against Double Deep Q-Network (DDQN), Power Follower Controller (PFC) and Fuzzy Logic Controller (FLC). • Using New York City cycle as a validation drive cycle, the approach improves fuel economy by 7.68%, 13.53%, and 10% compared to DDQN, PFC, and FLC, respectively. • The deep REINFORCE approach improves fuel economy by 5.31 %,9.78 %, and 9.93 % compared to DDQN, PFC, and FLC, respectively under another validation cycle, Amman cycle. • The proposed method has 38% less training time when compared to the DDQN approach. Fuel cell hybrid electric vehicles offer a promising solution for sustainable and environment friendly transportation, but they necessitate efficient energy management strategies (EMSs) to optimize their fuel economy. However, designing an optimal leaning-based EMS becomes challenging in the presence of limited training data. This paper presents a deep stochastic reinforcement learning based approach to address this issue of epistemic uncertainty in a midsize fuel cell hybrid electric vehicle. The approach introduces a deep REINFORCE framework with a deep neural network baseline and entropy regularization to develop a stochastic policy for EMS. The performance of the proposed approach is benchmarked against three EMSs: i) a state-of- art deep deterministic reinforcement learning technique called Double Deep Q-Network (DDQN), Power Follower Controller (PFC) and Fuzzy Logic Controller (FLC). Using New York City cycle as a validation drive cycle, the deep REINFORCE approach improves fuel economy by 7.68%, 13.53%, and 10% compared to DDQN, PFC, and FLC, respectively. The deep REINFORCE approach improves fuel economy by 5.31 %,9.78 %, and 9.93 % compared to DDQN, PFC, and FLC, respectively under another validation cycle, Amman cycle. Moreover, the training results show that the proposed algorithm reduces training time by 38% compared to the DDQN approach. The proposed deep REINFORCE-based EMS shows superiority not only in terms of fuel economy, but also in terms of dealing with epistemic uncertainty. [ABSTRACT FROM AUTHOR]

• A hierarchical predictive energy management is designed for fuel cell trucks under vehicle-following scenarios. • The speed planning considering future changes of leading vehicle is designed to ensure safe inter-vehicle distance and minimize power demand. • A driving-habit-conscious fuzzy C-means clustering enhanced Markov Chain speed prediction method is proposed to forecast future speed of leading vehicle. • The proposed hierarchical predictive energy management for host vehicle can reduce fuel cell degradation cost by 1.31–3.48% and total operation cost by 0.75–1.94%. Fuel cell hybrid electric heavy-duty vehicles play a key role to realize green and low-carbon travel. To cope with actual traffic environment, an advanced energy management strategy (EMS) is crucial to ensure vehicle's efficiency and economy. In this paper, a predictive co-optimization control method is designed to achieve speed planning and energy management for the host vehicle with a fuel cell/Li-ion battery hybrid energy storage system under vehicle-following scenarios. Firstly, the host vehicle obtains the real-time speed of leading vehicle by communication technology, and speed prediction for the leading vehicle is implemented by fuzzy C-means clustering enhanced Markov Chain (FCM-MC) considering driving habits. Secondly, based on speed prediction results, the speed planning which aims to ensure safe inter-vehicle distance and minimize the demanded vehicular power is implemented to derive the speed curve of host vehicle. Finally, according to speed planning results, predictive energy management is applied to achieve power allocation between fuel cell (FC) and battery. The simulation results denote that the proposed speed prediction method can decrease forecast error effectively. The speed planning ensures the safe inter-vehicle distance. With speed prediction accuracy improved, the proposed hierarchical energy management strategy can reduce fuel cell degradation cost by 1.31%-3.48% and total operation cost by 0.75%-1.94%. [ABSTRACT FROM AUTHOR]