Noise modeling of F22-A fighter jet using metaheuristic methods

Abstract The noise levels produced by high-performance supersonic military aircraft engines significantly exceed those of civilian aircraft, highlighting the critical importance of predicting military aircraft noise levels. This paper shows the modeling and assessment of the correlation between sound frequency and sound pressure level (SPL) using particle awarm optimisation (PSO) and the cuckoo search algorithm (CSA) for the high-performance fighter aircraft F-22 Raptor. The developed model aims to predict noise with high precision at various microphone angles from 60° to 150°. As a result of the analysis, the MAPE value for CSA was found to be below 1% for 10 different inlet angles, while the maximum mean absolute percentage error (MAPE) for PSO was 1.7863%. The large dataset ranging from 238 to 762 data points are used and the minimal error values confirm the high accuracy of the model. Additionally, the PSO and CSA algorithms were compared indirectly. The lower error values for CSA, along with its correlation coefficient (R) values closer to 1 indicate that the CSA method gives better results than PSO.