Sound pressure prediction at measurement points in high-intensity sound and flow-through pipeline

In aviation acoustic tests conducted in high-intensity sound and flow-through pipelines, real-time monitoring of sound pressure signals at designated measurement points is required to ensure test accuracy. However, direct microphone installation at these points significantly perturbs the vibration response of thin-walled pipeline structures. Therefore, it is necessary to predict sound pressure at critical measurement points using measurements acquired from non-critical locations. Most existing theories and methods rely on assumptions of linearity, quiescent (no-flow) conditions, and low-intensity sound environments—assumptions that are fundamentally inconsistent with the actual aviation acoustic testing scenarios addressed in this paper. To address the challenge of predicting surface sound pressure on test specimens under low-speed airflow and high-intensity sound pipeline conditions, this paper proposes an error backpropagation neural network–based sound pressure prediction algorithm to establish a time-domain mapping between measured and target point sound pressures. The experimental results demonstrate that the proposed method achieves a sound pressure prediction error as low as 0.01 dB in the critical frequency bands under the experimental conditions, and achieves better accuracy and lower computational complexity compared with the traditional model and the other two network models.