The invention provides a
turbine blade design method which comprises the following steps: S1, dividing a
turbine blade profile into four sections at equal intervals along the radial direction according to the size of a
turbine, and calculating section parameters of a middle-section
turbine blade profile; s2, connecting the four sections by adopting a third-order Bezier curve, defining a blade
front edge angle theta as an included angle between the Bezier curve and a blade
front edge radial inlet
angle alpha, and generating a three-dimensional blade model by changing the inclination angle theta of each section; s3, performing optimization to obtain an optimized theta parameter; and S4, according to the optimized theta parameter, fitting each section blade profile by using a third-order Bezier curve to generate a turbine
stator and rotor three-dimensional model. According to the method, the
machine learning method is applied to design of the
turbine blade, the three-order Bezier curve, the CFD numerical
simulation method and the
particle swarm optimization are utilized, the blade parameters are given, the geometric parameters can be quickly and accurately predicted, the calculated amount is reduced, the cost is reduced, the design efficiency is improved, and the quality of the designed blade is guaranteed.