Sensitivity Analysis And Robust Design For A Turbine Disc With Arbitrary Distribution Under Low Cycle Fatigue Failure

The turbine disk is a core component of Aero-engine, and its life determines the life span of the whole. Hence assessing the life of an aero-engine turbine disk has a significant science value. At present a myriads of experiments have shown that its main failure mode is low-cycle fatigue and the main method of low-cycle fatigue life is certainty analysis through the Manson-coffin formula which was amended by a variety of methods. However, the turbine disc mostly works under multi-axial load and then the traditional one-axial fatigue prediction theory tends to make out big errors. These years the multi-axial fatigue theory gets more and more research and the application. What is more, the geometry and material parameters of the turbine disk unavoidly contains uncertainty because of the error in manufacturing, installation and other processes. Therefore the life of turbine disk also exists uncertainty and then the low-cycle fatigue life assessment in certainty will not be able to meet the needs of practical engineering.In this paper, after summing up the fatigue life assessment of turbine disk material, the SWT multi-axial formula was randomized by analysizing the fatigue data of the GH4169turbine disc forgings.Then the relative fatigue life reliability model was set up, which considers the influence of the structural size, material and load’s randomness.In addition, In term of the introduction of Edgeworth Prgression Perturbation Method, this article carried on the failure probability analysis of the turbine disc with arbitrary distribution and then got the sensitivity of the analysis parameter. Furthermore, by carrying on the robust design which based on the sensitivity analysis, some design proposal was proposed. Finally, because the turbine is a system which is composed of72key ways, the full turbine disc’s failure probability was studied by considering the influence of the related failure.The paper mainly including of following three aspects:(1) The finite element simulation model for the Turbine disk in three-dimension was estabished. The stress distribution and possible failure areas were gained through the elastic-plastic analysis, which was the foundation for reliability analysis of the fatigue life.(2) According to the fatigue character of this turbine disc, the basic parameters of the strain-life curve were got by analysizing the fatigue data of the GH4169turbine disc forgings. What is more, some life preidiction models were introduced and then the relative life was gained. The compared result showed that the SWT formula was acuratest. Finally, the fatigue life reliability model was gained by randomizing the SWT formula.(3) Through the introduction of the Edgeworth Progress Perturbation Method, this article carried on the failure probability and its sensitivity analysis. Because the turbine disc’s structure is very complex, it is hard to get the explicit expression of the limit state function. In order to solve this problem, this article introduced the response surface model. The most commonly response surface model is response surface method, Krigring method and the RBF neutral network method. This article adapted these three methods and the comparison with the result of NESSUS showed that all these methods can be used to calculate this turbine disc’s fatigue reliability and the Kriging-Edgeworh method is best.(4) Based on the failure probability analysis under constant load amplitude, the failure probability probability analysis under the variable load amplitude. The relative failure probability was got by Kriging-MCS method and Kriging-Edgeworth method.(5) Based on the failure probability sensitivity, this article carried on the robust design and then proposed some design proposal.(6) Because the turbine is a system which is composed of72key ways, the full turbine disc’s failure probability was studied by considering the influence of the related failure. Then the influence of each key’s relevance to the reliability and the failure life was gain.