声子晶体梁边界态的强化学习设计方案与本征模式分析

Reinforcement Learning Design Scheme and Eigenmode Analysis of Edge States of Phononic Beams

摘要:

研究周期性阶梯截面声子晶体梁，在理论推导的基础上通过建立强化学习模型框架实现了预期带隙的结构设计功能，提出了从本征模式的角度确定边界态频率的方法，并分别通过对整体结构和半结构的透反射分析验证该方法的准确性。在此基础上，检验了所设计的边界态对弹性波的鲁棒传输特性。提出的强化学习设计方案和本征模式分析边界态的方法使声子晶体梁边界态的设计分析简单易行，有助于推动设计边界态以实现弹性波精准调控的研究。

关键词:声子晶体梁;弹性波;拓扑边界态;强化学习;鲁棒性

Abstract:

The periodic stepped section phononic crystal beam is studied. Based on the theoretical derivation， the structural design function of the expected band gap is realized by establishing the reinforcement learning model framework. A method to determine the edge state frequencies from the perspective of eigenmode is proposed， and the accuracy of this method is verified by the transmission and reflection analysis of the whole structure and semi structures， based on which， the robust transmission characteristics of the designed edge state for elastic waves are tested. The reinforcement learning design scheme and the eigenmode analysis method of edge states proposed make the design and analysis process of edge states of phononic beams simple， and help promote the research of designing edge states to realize the accurate regulation of elastic waves.

Key words:phononic beam;elastic wave;topological edge state;reinforcement learning;robustness