To accurately design the disk and blade-tip clearance of high pressure turbine, turbine disk radial deformation was analyzed from a probabilistic perspective. Extremum Response Surface Method (ERSM) with high efficiency and high precision was introduced and the mathematical model of ERSM was established for nonlinear dynamic probabilistic analysis. Fully considering the nonlinearity of material and boundary conditions, the dynamic of heat loads and rotor speeds and enough random parameters, the dynamic probabilistic analysis of turbine disk radical deformation was completed by using the ERSM. The distribution characteristics of input-output variables and important factors on the radial deformation were gained. Through the comparison of methods, the results show that the ERSM can greatly reduce the computing time and improve the computational efficiency while keeping the acceptable calculation precision, and it is a feasible and perfect method in the nonlinear dynamic probabilistic analysis. It is obvious that ERSM will provide a promising way to design and optimize turbine disk and blade-tip clearance more effectively in future.
