Axial flow cyclones are widely used to separate gas-particle mixtures in many industrial applications. This study aims to optimize the axial cyclone pressure drop, removal rate, and sound levels. Single and multi-objective optimizations using a surrogate model were used to minimize the Euler number, Stokes number, and overall sound pressure level (OASPL). A radial basis function artificial neural network approach was also applied to generate more reliable data. The performed three single-objective genetic algorithms optimization studies led to the minimized Euler, Stokes, and OASPL numbers values of 5.082, 0.333, and 3.607, respectively. Then, the multi-objective NSGA-II optimization technique was used to generate a Pareto front that decision-makers can use for optimal performance. It was found that the concave guide vane, the guide vane revolution, and swirler geometry significantly affect the performance of axial cyclones. The low values of acoustic noise were recorded at low tangential and axial velocities values.
