For plug-in hybrid electric vehicle (PHEV), using a hybrid energy storage system (HESS) instead of a single battery system can prolong the battery life and reduce the vehicle cost. To develop a PHEV with HESS, it is a key link to obtain the optimal size of the power supply and energy system that can meet the load requirements of a driving cycle. Since little effort has been dedicated to simultaneously performing the component sizing of PHEV and HESS, this paper proposes an approach based on the particle swarm optimization (PSO) algorithm to simultaneously determine the sizes of the engine, motor, battery and supercapacitor (SC) in a PHEV with HESS. The drivetrain cost is minimized in a different all-electric range (AER)—and depends on the battery type—while ensuring the driving performance requirements. In addition, the effects of the power system and drive cycle on the component sizes were analyzed and compared. The simulation results show that the cost of the PHEV drivetrain with the Ni-MH battery/SC HESS is reduced by up to 12.21% when compared to the drivetrain with the Li-ion battery/SC HESS. The drivetrain cost is reduced by 8.79% when compared to analysis-based optimization. The type of power supply system and drive cycle can significantly affect the optimization results.

In this paper, we propose an acoustic black hole radial elastic metamaterial (AREM). Through the study of its dispersion relations, it is found that, compared with the conventional elastic metamaterial, the AREM gathers energy at the tip of the black hole cell, which can trigger the local resonance (LR) effect and couples with the Bragg scattering (BS) effect, thus opening the very low-frequency strong attenuation broadband. The influence of the structural parameters of the AREM on the bandgap (BG) characteristics is further explored, and the bandwidth can be modulated in the frequency range of 0–1300 Hz by varying the truncation thickness and power exponent of the acoustic black hole (ABH) structure. Finally, by analyzing the transmission spectrum and displacement field, it is found that the total bandwidth of the flexural BG is better than that of the conventional radial elastic metamaterial, and the wave attenuation capability is improved by more than 110%. It is also discovered that the BG characteristics of the longitudinal BG are also better than those of the conventional radial elastic metamaterial, and the total bandwidth of the longitudinal BG is superior to that of the conventional radial elastic metamaterial, with the wave attenuation capability improved by more than 56%. The research findings may have applications in engineering fields such as ultra-low-frequency vibration reduction.

This paper presents a novel communication-less control method for PV-ESS DC microgrids. First, the ESS closest to the critical load is selected as the master unit, and output voltage is set at the nominal voltage. This unit injects an AC signal into the DC bus, and the frequency of the signal is determined by the unit's SOC and current. Then, other ESSs (called slave units) regulate their output currents according to the received AC signal frequency and local SOC without participating in the bus voltage regulation. In addition, thelogic signal is introduced to determine whether the PV module operates at maximum power point Tracking (MPPT) or droop control mode based on the received frequency. Simulation results demonstrate that the proposed control method achieves superior performance compared to traditional droop control. The method ensures higher load sharing accuracy, lower voltage deviation, and SOCs equalization in the DC microgrid system relying no communication link.

Purely data-driven approaches for remaining useful life (RUL) prediction of lithium-ion batteries often suffer from poor physical interpretability and limited generalization across heterogeneous cells. These limitations can lead to unreliable predictions and unsafe decisions in battery management systems. To overcome these challenges, this paper proposes a Physics-Informed Neural Network (PINN) framework that integrates a semi-empirical degradation ordinary differential equation (ODE) into the loss function, thereby enforcing strict physical monotonicity and eliminating non-physical oscillations. Furthermore, to mitigate distribution shifts caused by battery variability, an adaptive transfer learning strategy is developed. By parameterizing the physical decay rate as a trainable variable, the model achieves rapid adaptation to target batteries using minimal early-cycle data. Validated on the NASA and MOLICEL datasets, the proposed method ensures physical consistency and demonstrates superior accuracy (RMSE=0.0049) compared to state-of-the-art models. Ultimately, this approach effectively resolves the "cold start" problem in cross-battery deployment, offering a robust paradigm for reliable battery health management.