Key Laboratory of Underwater Acoustic Communication and Marine Information Technology, College of Ocean and Earth Sciences, Xiamen University, Xiamen 361010, People’s Republic of China
Auto-generated profile, based on publicly available information
Wuyi Yang
Key Laboratory of Underwater Acoustic Communication and Marine Information Technology, College of Ocean and Earth Sciences, Xiamen University, Xiamen 361010, People’s Republic of China
Resonant ultrasound spectroscopy (RUS) is an excellent alternative to electric resonance for characterizing the elastic and piezoelectric constants (EPCs) of piezoelectric materials with high mechanical quality factors. The RUS inversion tool must tolerate mode omissions because they are often difficult to avoid when using RUS measurements. In this study, a combination of RUS and a deep neural network (DNN) was proposed to characterize the EPCs of piezoelectric materials. The DNN configured to permit mode omission was designed and trained to precisely determine the EPCs from the resonance frequencies under the given clamped dielectric constants, density, and dimensions. The practicability and reliability of using the DNN as an inversion tool in RUS with mode omission tolerance were demonstrated by characterizing the EPCs of lead zirconate titanate ceramics. The DNN, after it was trained, was more efficient than the traditional nonlinear least squares method.
6 May 2025 in IEEE Transactions on Instrumentation and Measurement
Owing to their superior transverse electromechanical properties, [011]C poled relaxor-based ferroelectric single crystals are suitable for fabricating high-performance transverse-mode sensors and actuators. Device operation at various temperatures requires an understanding of the temperature-dependent full matrix material constants of [011]C poled relaxor-based single crystals. However, there are no published results regarding this yet, because the determination of these constants is complicated by the number of variables involved and the inconsistencies introduced by measuring multiple samples using conventional methods. In this study, the temperature-dependent full matrix material constants of [011]C poled Pb(Mg1/3Nb2/3)O3–28PbTiO3 single crystals were characterized using a single rectangular block sample. A modified resonant ultrasound spectroscopy technique was employed to determine the elastic and piezoelectric constants in the temperature range of 20–50 °C. To improve the mode identification efficiency and inversion reliability, the resonance frequencies of the sample under different boundary conditions were used during inversion. This approach overcame the self-consistency issue associated with results obtained using multiple-sample methods, such as electric resonance.