森林工程2024，Vol.40Issue(4)：160-167,8.DOI:10.7525/j.issn.1006-8023.2024.04.017
基于木材振动特性的月琴声学品质广义回归神经网络预测模型
Research on GRNN Prediction Model for Acoustic Quality of Yueqin Based on Wood Vibration Characteristics
摘要
Abstract
Paulownia has usually been an important material for making resonant components of musical instruments,which has a significant influence on the sound quality of musical instruments.This study utilized a generalized regression neural network(GRNN)to develop the sound quality evaluation model of Yueqin based on the vibration performance of the soundboard.In this study,nine Yueqins were fabricated,and a prediction model for the sound quality of Yueqins was proposed based on their sound qual-ity evaluation and the soundboard information of prepared Yueqins.Out of a total of 180 sets of data,135 sets of data were randomly selected for training and the remaining 45 sets of data were used for validation.A model for evaluating the acoustic quality of Yueqin instruments was established using principal component analysis method and GRNN,and simulation prediction was performed.The re-sults showed that based on the vibration characteristics of the soundboard,the prediction of the Yueqin sound quality can be achieved by using the Matlab simulation,and the accuracy of the prediction can reach 91.41%.In addition,this study demonstrated that the dynamic elastic modulus,acoustic radiation damping coefficient,elastic modulus,elastic and shear modulus ratio,acoustic imped-ance,loss tangent angle,and acoustic conversion efficiency of Paulownia wood resonator plates were all key factors influencing the acoustic quality of the finished Yueqin during its preparation.关键词
广义回归神经网络/主成分分析/声学品质/振动特性/共鸣板/木材/民族乐器Key words
Generalized regression neural network/principal component analysis/acoustical quality/vibration characteristics/soundboard/wood/national musical instrument分类
农业科技引用本文复制引用
杨扬..基于木材振动特性的月琴声学品质广义回归神经网络预测模型[J].森林工程,2024,40(4):160-167,8.基金项目
中国博士后科学基金项目面上项目(2019M651240). （2019M651240）
