Physics of failure-based reliability prediction of turbine blades using multi-source information fusion

Abstract Fatigue and fracture of turbine blades are fatal to aero engines. Reliability prediction of aero engines is indispensable to guarantee their safety. For turbine blades of aero engines, most recent research works only focus on the number of cycles and excavate information from a single source. To remove these limitations, a Physics of failure-based reliability prediction method using multi-source information fusion has been developed in this paper to predict the reliability of turbine blades of aero engines. In the proposed method, the fuzzy theory is employed to represent uncertainties involved in prediction. Case studies of reliability prediction under fuzzy stress with and without fuzzy strength are conducted by using a dynamic stress-strength interference model which takes types of cycles of aero engines into consideration. Results indicate that the proposed method is better in line with engineering practice and more flexible in decision making and it can predict the reliability of aero engine turbine blades to be an interval by utilizing the proposed linear fusion algorithm. In addition, the predicted interval contains results that are predicted by other commonly used information fusion methods Hence, the proposed method conduces to remove confusion made by selection of multiple methods.