A symmetrical one-dimensional(1D)photonic crystal structure with a Dirac-emimetal-defected layer is proposed.The material properties of the Dirac semimetal are governed by three key parameters:Fermi level,Fermi veloci...A symmetrical one-dimensional(1D)photonic crystal structure with a Dirac-emimetal-defected layer is proposed.The material properties of the Dirac semimetal are governed by three key parameters:Fermi level,Fermi velocity,and degeneracy factor.Simulation results demonstrate that the proposed structure generates multiple photonic bandgaps within the THz frequency range.In the low-THz region,pronounced resonant transmission peaks emerge,enabling near-perfect filtering performance.The positions of these defect modes can be dynamically tuned by adjusting the Fermi level and degeneracy factor.In mid-and high-THz frequency bands,the Dirac semimetal begins to exhibit metallic behavior,leading to attenuation of the transmission peaks and the appearance of absorption.The elevation of the Fermi level delays the critical threshold for the transition from the dielectric state to the metallic state,while an increase in Fermi velocity suppresses metallic behavior.Therefore,enhancing both the Fermi level and Fermi velocity contributes to strengthening the defect peak intensity.Conversely,increasing the degeneracy factor strengthens the metallic characteristics,thereby disrupting the high-frequency photonic bandgap.Notably,the defect layer thickness and incident angle exert significant influence on the transmission behavior:a larger incident angle causes the defect peak to shift toward higher frequencies and reduces its intensity,whereas a thicker defect layer shifts the defect peak toward lower frequencies.The modulation effects of both parameters become more pronounced as frequency increases.Compared with conventional photonic crystals,our work can provide a tunable structure over transmission properties,offering novel strategies for designing tunable filters and optical sensors.展开更多
文摘A symmetrical one-dimensional(1D)photonic crystal structure with a Dirac-emimetal-defected layer is proposed.The material properties of the Dirac semimetal are governed by three key parameters:Fermi level,Fermi velocity,and degeneracy factor.Simulation results demonstrate that the proposed structure generates multiple photonic bandgaps within the THz frequency range.In the low-THz region,pronounced resonant transmission peaks emerge,enabling near-perfect filtering performance.The positions of these defect modes can be dynamically tuned by adjusting the Fermi level and degeneracy factor.In mid-and high-THz frequency bands,the Dirac semimetal begins to exhibit metallic behavior,leading to attenuation of the transmission peaks and the appearance of absorption.The elevation of the Fermi level delays the critical threshold for the transition from the dielectric state to the metallic state,while an increase in Fermi velocity suppresses metallic behavior.Therefore,enhancing both the Fermi level and Fermi velocity contributes to strengthening the defect peak intensity.Conversely,increasing the degeneracy factor strengthens the metallic characteristics,thereby disrupting the high-frequency photonic bandgap.Notably,the defect layer thickness and incident angle exert significant influence on the transmission behavior:a larger incident angle causes the defect peak to shift toward higher frequencies and reduces its intensity,whereas a thicker defect layer shifts the defect peak toward lower frequencies.The modulation effects of both parameters become more pronounced as frequency increases.Compared with conventional photonic crystals,our work can provide a tunable structure over transmission properties,offering novel strategies for designing tunable filters and optical sensors.
文摘针对不同磁密幅值、频率、谐波组合等复杂激励工况下磁致伸缩建模面临的精准性问题,该文利用空间注意力机制(spatial attention mechanism,SAM)对传统的卷积神经网络(convolutional neural network,CNN)进行改进,将SAM嵌套入CNN网络中,建立SAMCNN改进型网络。再结合双向长短期记忆(bidirectional long short-term memory,BiLSTM)网络,提出电工钢片SAMCNN-BiLSTM磁致伸缩模型。首先,利用灰狼优化算法(grey wolf optimization,GWO)寻优神经网络结构的参数,实现复杂工况下磁致伸缩效应的准确表征;然后,建立中低频范围单频与叠加谐波激励等复杂工况下的磁致伸缩应变数据库,开展数据预处理与特征分析;最后,对SAMCNN-BiLSTM模型开展对比验证。对比叠加3次谐波激励下的磁致伸缩应变频谱主要分量,SAMCNN-BiLSTM模型计算值最大相对误差为3.70%,其比Jiles-Atherton-Sablik(J-A-S)、二次畴转等模型能更精确地表征电工钢片的磁致伸缩效应。
