Optimization of the Laminar Flow Fan Blade Profiles Based on Artificial Neural Network

Alaminar flow optimization scheme for fan blade profiles is studied. The first step is to parameterize the original profiles with the class/shape function transformation (CST) approach that provides accurate definitions of the blade profile geometries. Then the main aerodynamic performances of the profile (including the pressure ratio, the efficiency and the size of laminar flow regions on the profiles) are evaluated by CFD tools. By repeating the above steps, a set of profile geometry parameters and their relevant aerodynamic performances are obtained. With the help of artificial neural network, the relations between the geometric parameters and the aerodynamic performances can be built, which can be used to re-define the geometry of the fan bade profile. By applying the scheme above, the profiles on different relative spanwise positions of the fan blade are optimized. The results show the increases of either the efficiency or the size of the laminar flow regions on the optimized blade profiles