In order to improve the reliability analysis precision of turbine blades

the fuzzy response surface method (FRSM

Fuzzy Response Surface Method) of structural reliability analysis is put forward considering the fuzziness of input variables and limit states. First

using the method of equivalent transformation

the input variables with fuzziness is converted into equivalent random variables. Then

the mathematic model of fuzzy response surface for structural reliability analysis is established based on the quadratic polynomial response surface function. The mean value and the variance of the stresses and radial deformations of the blades were gained through a lot of joint sampling of the model using the Monte Carlo method. Finally

the event failure probability and the fuzzy random reliability index are calculated based on the probability integral method. The analysis result shows that the blade reliability indexes decrease with the increase of fuzzy coefficient. Compared with the traditional method based on random variables

this method is more reasonable for the description of uncertainty

and the evaluation of reliability is more accurate theoretically.