ItemArticle

Hybrid Analytical-Machine Learning Framework for SH-Wave Dispersion in Piezo-Flexoelectric Layered Structures with Imperfect Interfaces

Title
Hybrid Analytical-Machine Learning Framework for SH-Wave Dispersion in Piezo-Flexoelectric Layered Structures with Imperfect Interfaces
Creator
Seema; Singhal, Abhinav
Description
This study presents a hybrid analytical-machine learning (ML) framework for modeling shear-horizontal (SH) wave propagation in piezo-flexoelectric (PFE) layered structures with imperfect interfacial bonding. The governing dispersion relations were derived using mechanical, electrical, and flexoelectric continuity conditions, establishing explicit links among phase velocity, wave number, layer thickness ratio, and flexoelectric coefficients. Analytical results revealed that nanoscale flexoelectricity and interfacial imperfections significantly influence wave behavior: electrically open (EO) boundaries amplify electromechanical coupling and enhance dispersion, whereas electrically shorted (ES) conditions suppress these effects, leading to reduced phase velocities. To address the computational cost associated with evaluating high-dimensional parametric spaces, ML surrogate models were incorporated. Convolutional neural network (CNN) and k-nearest neighbor (KNN) regressors accurately reproduced analytical dispersion surfaces with errors below 3% and drastically reduced computation time by more than two orders of magnitude. Classification models provided >90% accuracy in distinguishing EO and ES boundary-condition regimes, and generative artificial intelligence (AI) variational autoencoder/generative adversarial network (VAE/GAN) architectures successfully synthesized dispersion surfaces for previously unseen flexoelectric parameter ranges. The proposed hybrid framework combines the physical rigor of analytical modeling with the efficiency of modern ML surrogates, enabling rapid parametric exploration and high-fidelity prediction of SH-wave characteristics. The outcomes support accelerated design of advanced surface acoustic wave (SAW) devices, micro-electro-mechanical systems (MEMS) components, and nanoscale electromechanical systems. Future work will extend the approach to experimental validation and uncertainty-aware modeling for real-world applications. 2026 American Society of Civil Engineers.
Source
Journal of Engineering Mechanics;Volume;152;Issue;6;Article No.;4026020;
Date
01-01-2026
Publisher
American Society of Civil Engineers (ASCE)
Subject
Imperfect interfaces; Machine learning surrogates; Physics-informed neural networks; Piezo-flexoelectricity; Shear-horizontal waves; Surface acoustic wave (SAW)/micro-electro-mechanical systems (MEMS) applications
Coverage
Seema, Dept. of Mathematics, Christ Univ., Bengaluru, Karnataka, 560029, India; Singhal A., Dept. of Mathematics, Christ Univ., Bengaluru, Karnataka, 560029, India
Rights
Restricted Access; Hardcopy may be available in the library
Relation
ISSN: 7339399;
Format
online
Language
English
Type
Article

Citation

Seema; Singhal, Abhinav, “Hybrid Analytical-Machine Learning Framework for SH-Wave Dispersion in Piezo-Flexoelectric Layered Structures with Imperfect Interfaces,” CHRIST (Deemed To Be University) Institutional Repository, accessed June 18, 2026, https://archives.christuniversity.in/items/show/22616.

Output Formats