Flexoelectricity is an electro-mechanical coupling effect between electric polarization and mechanical strain gradient. Dielectric materials, such as ceramics, rubber, plastics, etc., have a flexoelectric effect, and bending or twisting to them will generate electricity, but not when applying pressure or tension, which greatly limits its application, because in fact tension and pressure loads are the easiest to obtain. In this work, we establish a theoretical model and asymmetric structure in polymer elatomers, in which complex asymmetric structures can exhibit flexoelectric effects under any load due to their local structural asymmetry compared to traditional structures. By comparing a series of parameters of the asymmetric structure, we design a structure in elastomer, as a result, its equivalent piezoelectric coefficient is much larger than the material flexoelectric coefficient. The results show that the polymer elastomers with asymmetric structure can be widely used in the energy harvesting, sensing and mechanical actuating with high flexoelectric effect.

In space optical applications, the piezoelectric-actuated fast steering mirror (FSM) is one of the pivotal components for high-precision beam capturing and trajectory tracking. The FSM is restrained in small-angle scanning applications due to the short actuation stroke of the incorporated piezoelectric materials. This study introduces a dual-axis sub-radian stroke FSM with high precision and self-sensing capability, based on cascading structures for displacement amplification. Experimental results reveal that both axes can rotate 148.67 mrad, comparable to existing electromagnetic FSMs.

Bridge arm reactors are subjected to composite current of AC and DC in operation. Aiming to study the differences in temperature rise, current and loss calculations are performed according to electromagnetism and heat transfer theories. Considering the influence of metal accessories, the temperature distribution of the reactor and its influencing factors are analyzed. The results show that the loss of metal accessories accounts for 1.02% and the collector ring is prone to local overheating. The maximum temperature rise of the encapsulation shows a nonlinear decrease with the increase of the flow rate and the decrease of the ambient temperature. The maximum temperature rise is increased by 14.6℃ after considering the sound insulation device. Finally, the error is 5.72% when comparing the results of AC and DC temperature rise tests, which verifies the correctness of the simulation model. The research results can provide reference for the design of temperature rise of reactors under AC-DC compound high current in actual engineering.

Physics-informed neural networks (PINNs), which is a powerful approach for solving partial differential equations with deep learning, has been recently applied to modeling of electrodynamic interaction problems including a relativistic beam in charged particle accelerators. In the present study, the transfer learning (TL) is applied to the PINNs based on the total-field (TF) formulation. It is shown that TL can accelerate significantly the training process of the TF-PINNs in the simulation of the beam impedance of an infinitely long beam pipe of circular cross section.

To address the inconvenience and the maintenance cost of conventional batteries in the development of large-scale low-power wireless sensors, radio frequency (RF) energy harvesting (EH) technology has gained significant tractions since RF energy offers advantages such as abundant availability and resilience to environmental conditions. In this study, we propose and optimize a novel topology for a quad-band and polarization-insensitive metasurface (MS) absorber. The MS unit employs four spoof-local-surface-plasmons (SLSPs) resonators arranged using a sequentially-rotation technique, enabling a polarization insensitivity and facilitating a post-stage power accumulation. We then develop an optimization methodology by combining a hybrid algorithm and finite element method simulations to calculate the performance parameters. The methodology is applied to a prototype design, demonstrating exceptional results.

Millmeter wave band electromagnetic waves are used in 5G communication systems. They have a wide frequency range for wireless communication, however their short wavelength makes them susceptible to relatively small obstacles such as robots and people. In this study, we use machine learning to estimate the wave propagation of electromagnetic waves in the 28 GHz band in order to estimate the effect of obstacles. The ray tracing method was used as the teacher data for the machine learning. Comparison of the machine learning estimation results with the experimental results shows good agreement.

We synthesize non-spherical magnetic particles and evaluate their heat generation efficiency under an ac magnetic field. Ferrocene is dissolved in ethanol and the solution is heated up to reach a supercritical state. Spherical, dendritic and flower-shaped magnetic particles are synthesized under the supercritical condition and their production rates change depending on the heating time. We show that the non-spherical magnetic particles exhibit higher heat generation efficiency under an ac magnetic field as compared to the spherical ones.