Structural Stability and Dynamic Behavior of Piezoelectric Plates in Landscape Design Assessed via Physics-Informed Deep Neural Networks

The research presents a new computational system which allows researchers to examine piezoelectric plate structural stability and dynamic performance in landscape design projects that need to withstand wind loads and use energy-harvesting systems. The researchers create wind conditions for their research by using two atmospheric parameters which include the wind attack angle and the average wind speed to simulate outdoor conditions that affect piezoelectric plate performance. The plates are assumed to rest on viscoelastic–concrete auxetic foundations which provide landscape-integrated structures with improved damping abilities and unique deformation patterns. A power-law material model enables the creation of piezoelectric plate material properties which simulate functional gradients that exist throughout the thickness of the material. The research uses higher-order shear deformation theory (HSDT) to establish a precise method for examining transverse shear effects which does not require shear correction factors. The electromechanical coupling mechanisms find their complete description through Maxwell’s equations, while Hamilton’s principle serves as the basis for deriving the equations which govern motion because it ensures energy conservation. The harmonic differential quadrature method uses Chebyshev polynomial grid points to achieve spatial discretization of coupled partial differential equations which results in high numerical accuracy and efficient computational performance. The primary solution method uses Physics-informed deep neural networks (PIDNNs) as an Artificial intelligence technique to address the deficiencies which traditional numerical solvers experience during complex multiphysics simulations. The PIDNN framework enables the prediction of stability boundaries and vibration characteristics, plus electromechanical responses through its integration of governing physical laws into the loss function. The proposed method successfully analyzes smart piezoelectric plate systems which operate in wind-sensitive landscape design applications.