Sound Absorption Optimization of Porous Materials Using BP Neural Network and Genetic Algorithm

In recent years, the interior noise of automobile has been becoming a significant problem. In order to reduce the noise, porous materials have been widely applied in automobile manufacturing. In this study, the simulation method and optimal analysis are used to determine the optimum sound absorption of polyurethane foam. The experimental simulation is processed based on the Johnson-Allard model. In the model, the foam adheres to a hard wall. The incident wave is plane wave. The function of the model is to calculate the noise reduction coefficient of polyurethane foam with different thickness, density and porosity. The back propagation neural network coupled with genetic optimization technique is utilized to predict the optimum sound absorption. A developed back propagation neural network model is trained and tested by the simulation data. By comparing the test data and the prediction data of neural network, it is obvious that back propagation neural network model is reliable and efficient to predict the sound absorption of polyurethane foam. The genetic algorithm is employed to identify the optimum sound absorption. The thickness, density and porosity are regarded as the inputs of genetic algorithm. The prediction result of the back propagation neural network is utilized as individual fitness value. The outputs of genetic algorithm are the optimum parameters of polyurethane foam. Then, an additional simulation is utilized to calculate the optimum noise reduction coefficient as 0.62.