基于多元回归方法的隔音垫隔音性能研究

Sound Insulation Performance Research of Sound Insulation Pads Based on Multiple Regression Method

摘要: 性能开发是产品研发中的核心环节，通常采用DoE的方式，筛选出影响因子并建立数学模型，以求解最佳参数值。DoE需要严格的实验设计，设定因子的高低水平组合，各处理的试验结果才可用于因子的主效应和交互效应等的估计。而多元回归方法的要求就没有那么多，只要引入足够多的变差，建立完整的性能响应模型，就可以逐步回归的方式剔除无效因子，获得理想的数学模型，进而可以求解满足性能要求的各因子水平。隔音性能是隔音垫的核心性能，预估的影响因子包括厚度、密度和形状。通过多元回归分析得到的响应数学模型，方便设计人员配置参数，极大方便了产品设计，有效提高设计通过性。

Abstract: Performance development is a critical component of product development, typically employing DoE to identify influencing factors and establish mathematical models to determine the optimal parameter settings. DoE requires meticulous experimental design, specifying the combinations of high and low levels of factors, so that the results from each treatment can be used to estimate the main effects and interactions of the factors. In contrast, multiple regression methods have fewer requirements; by introducing sufficient variations and establishing a comprehensive performance response model. The method can eliminate ineffective factors through gradually regression, leading to an ideal mathematical model that can determine the optimal parameters’ levels for meeting performance criteria. Sound insulation performance is the core attribute of soundproofing pads, with key influencing factors including thickness, density, and shape. The response mathematical model derived from multiple regression analysis facilitates parameter configuration by designers, greatly simplifying product design and enhancing design feasibility.