Abstract:Long Short-Time Memory (LSTM) deep neural networks are capable of learning order dependence in sequence problems and capturing long-term, non-linear temporal dependencies between the input and out of a system. With the long-term vision to model dynamical systems to which analytical or numerical methods are impossible or difficult to apply, this paper presents a study of modeling system dynamics and predicting responses using the LSTM networks, which have demonstrated excellent capability in predicting single-mode responses in a prior study. However, the LSTM network exhibits difficulties in modeling and predicting multi-mode responses accurately. To resolve the multi-mode issue, this paper presents an approach that obtains an equivalent network consisting of a set of sub-networks learned on isolated modes, and demonstrates its effectiveness on a simulated 2-degree-of-freedom (2DOF) mass-spring-damper system of nonlinear Duffing springs. The second part of the paper is focused on the application of the proposed approach in piezoelectric energy harvesting. Experiments are conducted on a harvester subjected to random base-motion excitation and exhibiting nonlinearity in its multi-mode response. Both the direct and mode-separation LSTM modeling approaches are applied to predict the output voltage given a random base-motion excitation. The mode-separation approach outperforms the direct approach significantly, and yields an excellent match between the actual and predicted responses. Specifically, for a test electrical voltage response of RMS value 0.2241 V, the difference between the actual test and predicted responses by using the mode-separation approach has an RMS value of 0.0504 V, compared to 0.1645 V obtained by using the direct LSTM approach. It is also much lower than the RMS value of 0.1835 V obtained by using the attention-based LSTM network, another comparison method. Leveraging a deep learning strategy, the validated approach opens up opportunities for accurately modeling energy harvesting systems of high complexities and/or strong nonlinearities.

Novel design of recurrent neural network for the dynamical of nonlinear piezoelectric cantilever mass–beam model

Modelling of Nonlinear Dynamics for Reciprocal Multi-Layer Piezoceramic Transducer Systems

Active vibration control of a cantilever beam based on neural network online identification

Research on Nonlinear Model of Piezoelectric Actuator

Nonlinear Modeling of a Piezoelectric Actuator-Driven High-Speed Atomic Force Microscope Scanner Using a Variant DenseNet-Type Neural Network

Modeling and Analysis of Cantilever Beam with Active-Passive Hybrid Piezoelectric Network

Long short-term memory (LSTM) neural networks for predicting dynamic responses and application in piezoelectric energy harvesting

Physics-informed neural network with data-driven in modeling and characterizing piezoelectric micro-bender

Modal Voltages of Linear and Nonlinear Structures Using Distributed Artificial Neurons

Bayesian model class selection of nonlinear constitutive relationships for piezoelectric energy harvesters with small set of observations

The artificial neural network modelling of the piezoelectric actuator vibrations using laser displacement sensor

Memory-Enhanced Neural Network Control of Piezoelectric Actuators With a Rate-Amplitude-Dependent Hysteresis Model

Additive Cascaded Model Identification Of Linear And Nonlinear Dynamics Of Piezoelectric Actuator

Study on ECNLP Dynamics Model of Piezoceramic Actuator and Position Tracking Controller

A Recurrent Neural Network-Based Method for Dynamic Load Identification of Beam Structures

A Consistent Geometrically Nonlinear Model of Cantilevered Piezoelectric Vibration Energy Harvesters

Modeling and Control of a Linear Piezoelectric Actuator

Analysis and Modeling of a Novel Piezo-Actuated Electromechanical Coupling Dynamic System

RKHS Embedding for Estimating Nonlinear Piezoelectric Systems

Programmable piezoelectric phononic crystal beams with shunt circuits: A deep learning neural network-assisted design strategy for real-time tunable bandgaps

Nonlinear Vibrations of the Piezoelectric-Laminated Composite Cantilever Rectangular Plate Subjected to the Transverse and Parametric Excitation