This work presents an efficient procedure to predict the sound reduction index (SRI) of a flat panel of arbitrary internal complexity starting from numerical finite element (FE) simulations. This hybrid analytical-FE procedure allows to perform narrow band calculations considering fluid coupling up to high frequencies. The exciting sound field is synthesised considering a series of uncorrelated plane waves impinging on the panel. A stochastic approach based on the sinc function is also considered. A test case is presented in which the hybrid approach is applied to a glass panel. Computation time and memory consumptions are compared for the different implementations, showing that the stochastic approach requires considerably larger resources than the plane wave formulations. The method is validated against predictions by analytical methods and experimental data from sound transmission suites. To match the experimental results, the angle of incidence is limited to the value of 75${}^{\circ },$ whose origin is discussed by means of ray-tracing simulations of the sound field distribution in the source room. It is found that the quality of the predictions can be improved by including the actual characteristics of the sound field in the plane wave formulation. If compared to a commercial full FE software, the proposed hybrid procedure features greatly reduced computational time, and results closer to the experimental values in the entire frequency range of interest.

这项工作提出了一种有效的程序，可以从数值有限元 (FE) 模拟开始预测具有任意内部复杂性的平板的降噪指数 (SRI)。这种混合分析有限元程序允许执行窄带计算，考虑到高频的流体耦合。考虑一系列不相关的平面波撞击面板，合成令人兴奋的声场。还考虑了基于 sinc 函数的随机方法。展示了一个测试案例，其中将混合方法应用于玻璃面板。比较了不同实现的计算时间和内存消耗，表明随机方法需要比平面波公式大得多的资源。该方法通过分析方法和来自声音传输套件的实验数据的预测得到验证。为了匹配实验结果，入射角被限制在 75 的值${}^{\circ },$其起源是通过声源室声场分布的射线追踪模拟来讨论的。发现可以通过在平面波公式中包含声场的实际特征来提高预测的质量。如果与商业完整 FE 软件相比，所提出的混合程序的特点是大大减少了计算时间，并且在整个感兴趣的频率范围内结果更接近实验值。