基于實(shí)時(shí)數(shù)據(jù)投影的動(dòng)態(tài)模態(tài)重構(gòu)方法及其在公路隧道內(nèi)聲場(chǎng)預(yù)測(cè)中的應(yīng)用

打開(kāi)文本圖片集

中圖分類(lèi)號(hào)：0213.6 文獻(xiàn)標(biāo)志碼：A 文章編號(hào)：1003-5168(2025）12-0070-05

DOI: 10.19968/j.cnki.hnkj.1003-5168.2025.12.013

Dynamic Modal Reconstruction Method Based on Real-time Data Projection and Its Application in Tunnel Acoustic Field Prediction

DONG Zhenlin (Qinghai Provincial Transportation Construction Engineering Cost Station,Xining8171O1,China)

Abstract: [Purposes]To address the unsuitability of existing modal reconstruction methods for predicting non-stationary physical fields with increasing prediction steps,this paper aims to enhance the adaptability of Dynamic Mode Decomposition (DMD) for non-stationary physical field prediction.[Methods] Based on Koopman-mode decomposition and data driven physical field reconstruction,this paper proposes a dynamic modal reconstruction method based on real-time data projection.Its effectiveness has been verified through two cases: strong sound-absorbing walls and highly reflective walls in a tunnel. [Findings] Unlike traditional methods that perform reconstruction using the initial physical field,the dynamic modal reconstruction method based on real-time data projection can effctively eliminate the error accumulation caused by the non-stationary acoustic field by continuously incorporating the latest real - time data for data mapping,thus significantly improving prediction accuracy.The proposed method can be an efficient prediction tool for non-stationary physical fields.[Conclusions] The error-accumulation effect of existing modal reconstruction methods originates from thecontinuous introduction of new information in non-stationary physical field data.Real-time data projection can resolve the erroraccumulation phenomenon.

Keywords: data-driven; non-stationary; acoustic signal; tunnel; dynamic mode decomposition

0 引言

近年來(lái)，隨著我國(guó)公路網(wǎng)絡(luò)的不斷完善，隧道的建設(shè)工程數(shù)量不斷增加。（剩余10252字）