Fatigue Life Prediction And Reliability Analysis Of Critical Aircraft Engine Rotating Components

Fatigue reliability, a powerful toolkit for structural reliability improvement of mechanical system, has now become a cutting-edge of research and a hot topic in reliability engineering. It has been studied extensively both by reliability engineers and academic researchers. As the ever increasing complexity of modern mechanical equipment such as aircraft engines, and the continual outbreak of fatigue failure as well as the widespread existing of uncertainties in real engineering, various theories and methods of fatigue reliability have been proposed to handle different uncertainties in the fatigue analysis and design of critical aircraft engine rotating components. Effects of these uncertainties on the fatigue life of rotating components can be clearly identified and improved. As a result, mechanical designs incorporated with fatigue reliability engineering naturally give rise to critical rotating components with high reliability, low cost, and robust performance, which provides a feasible basis for ensuring the safety and reliability of aircraft engine.The issue of uncertainty in real engineering has been mainly categorized into two major groups based on the characteristics of different uncertainties, i.e. the aleatory uncertainty and the epistemic uncertainty. Nowadays, fatigue reliability considering aleatory uncertainty has been studied a lot with abundant theories and methods and corresponding theoretical system has been basically established. However there is little work on fatigue reliability considering epistemic uncertainty. It still deserves further investigation.This paper is devoted to the issues in fatigue reliability by considering both aleatory uncertainty and epistemic uncertainty. Specific attention is laid on the fatigue reliability analysis of critical rotating components of an aircraft engine, such as turbine disks, turbine and compressor blades. In this paper, considering the effects of aleatory and epistemic uncertainties to fatigue life, a comprehensive investigation of fatigue life prediction and fatigue reliability analysis for critical rotating components of aircraft engines is presented. Various theories and methods are incorporated coherently to deliver a theoretically-sound and well-constructed study, including the theory of Copula, the method of fuzzy risk priority number(FRPN), the method of saddlepoint approximation(SPA), the method of maximum entropy, and the theory of fatigue cumulative damage.The contributions of this dissertation are summarized as follows:(1) Fuzzy risk priority number analysis for the turbine and compressor blades is proposed under considering the dependences of the multiple failure modes. Copula theory is incorporated to model the dependence among multiple failure modes. Aiming at the problem of multiple correlated failure modes within the risk priority number assessment, fuzzy theory is used to describe and integrate epistemic uncertainty within the reliability analysis. An integration of these methods into classical failure mode, effects and criticality analysis(FMECA) and fault tree analysis(FTA) is implemented. It is demonstrated good efficiency and credibility for real engineering application through case study.(2) A saddlepoint approximation-based method for fatigue reliability analysis is developed. A low cycle fatigue life prediction model is constructed for the turbine disk alloy. The SPA method is implemented for fatigue reliability analysis of the turbine disk alloy GH4133 under considering the effects of various uncertainties on components’ fatigue life for both the explicit and implicit life prediction models. It is essential for materials selection, design and ensuring high reliability of products. Illustrative examples have demonstrated that SPA is very apt for the fatigue reliability analysis of turbine disk with a large number of random variables or high nonlinearity of performance functions. It requires only a small number of samples without any distribution assumptions for random variables which meets the actual engineering examples better with low cost and high robust performance.(3) Developing a maximum entropy theory based method for fatigue crack growth life reliability analysis. A modified generalized passivation-lancet model for long fatigue crack propagation rate(GPLFCPR) is developed based on the crack closure principle and is applied to crack growth life prediction. The maximum entropy method is applied to reliability analysis of fatigue crack growth life for direct aging GH4169 turbine disks superalloy by considering the uncertainty of random variables in fatigue crack growth life model. Simulation results have shown that the maximum entropy theory is available for fatigue reliability analysis of crack growth life for turbine disk with high efficiency and precision.(4) Developing a Morrow generalized ? ?N curved surface model for fatigue life prediction and fatigue cumulative damage analysis of turbine disks. To account for the effect of cyclic stress amplitude a? and mean stress m? on fatigue life, the power fuction cyclic ? ?N curve with three parameters is integrated with the Morrow constant life curve. It then gives rise to an integrated model that describes the relationship between the stress amplitude a?, mean stress m? and the fatigue life N. Then, the generalized Morrow fatigue ? ?N curved surface model is constructed. The fatigue life is then assessed using this generalized model. Total estimation of fatigue life is performed using the aforementioned model. Results have shown that the fatigue life predicted by the combination model of three parameters power function equation and Morrow constant life curve is close to the rated life.