The largest single cause of component failures in modern military aircraft gas turbine engines is high cycle fatigue (HCF), exceeding the number attributed to low cycle fatigue, corrosion, overstress, manufacturing processes, mechanical damage, and materials. The HCF problem is a pervasive one, affecting all engine sections and a wide range of materials. In addition to the impact on engine component reliability, HCF problems cause significant economic impacts through field inspection and maintenance actions, and reduced readiness reliability.

High cycle fatigue in aircraft gas turbines—an industry perspective

An investigation of fatigue failures of turbine blades in a gas turbine engine by mechanical analysis

Materials in rotating machinery are typically subjected to vibratory loading from a number of sources which, in turn, is superimposed on mean stresses which result primarily from steady-state centrifugal loads. In addition, components subjected to vibratory stresses can sustain damage during manufacturing, break-in cycles, or during service such as from foreign objects, fretting, or other types of wear. The combination of vibratory and ‘steady’ stress levels can, for certain load levels, produce low cycle fatigue damage in addition to the damage produced from the high frequency (HCF) vibratory loading since the ‘steady’ stresses are actually low cycle fatigue (LCF) which results in one cycle for every startup and shutdown operation. Design for HCF is generally based on a Goodman diagram which takes into account the vibratory as well as the steady stress amplitudes for fatigue runout or fatigue under a given number of cycles. It does not, however, take into account the combined effects of LCF and HCF. In this investigation, the combined effects are demonstrated analytically by numerical examples which consider both the initiation and propagation phases of fatigue. In addition to the analysis of LCF/HCF interactions, considerations which must be accounted for in design are reviewed in light of a number of failures of components in service in U.S. Air Force fighter engines. A critical assessment of the concepts embedded in the use of the Goodman diagram is presented. Comments on the limitations on the use of a Goodman diagram for design are provided. Some suggestions are offered for the improvement of the design methodology for HCF which involve both damage tolerance considerations and methods for assessing and improving the margin of safety.

On the use of the Goodman diagram for high cycle fatigue design

Failure analysis of gas turbine rotor blades

Variation in the stability of a rotating blade disk with a local crack defect

A novel fracture mechanics approach has been used to predict crack propagation lives in gas turbine engine blades subjected to vibratory high cycle fatigue (HCF). The vibratory loading included both a resonant mode and a nonresonant mode, with one blade subjected to only the nonresonant mode and another blade to both modes. A life prediction algorithm was utilized to predict HCF propagation lives for each case. The life prediction system incorporates a boundary integral element (BIE) derived hybrid stress intensity solution, which accounts for the transition from a surface crack to corner crack to edge crack. It also includes a derivation of threshold crack length from threshold stress intensity factors to give crack size limits for no propagation. The stress intensity solution was calibrated for crack aspect ratios measured directly from the fracture surfaces. The model demonstrates the ability to correlate predicted missions to failure with values deduced from fractographic analysis. This analysis helps to validate the use of fracture mechanics approaches for assessing damage tolerance in gas turbine engine components subjected to combined steady and vibratory stresses.

Damage Tolerance Based Life Prediction in Gas Turbine Engine Blades Under Vibratory High Cycle Fatigue

A novel fracture mechanics approach has been used to predict crack propagation lives in gas turbine engine blades subjected to vibratory high cycle fatigue (HCF). The vibratory loading included both a resonant mode and a non-resonant mode, with one blade subjected to only the non-resonant mode and another blade to both modes. A life prediction algorithm was utilized to predict HCF propagation lives for each case. The life prediction system incorporates a boundary integral element (BIE) derived hybrid stress intensity solution which accounts for the transition from a surface crack to corner crack to edge crack. It also includes a derivation of threshold crack length from threshold stress intensity factors to give crack size limits for no propagation. The stress intensity solution was calibrated for crack aspect ratios measured directly from the fracture surfaces. The model demonstrates the ability to correlate predicted missions to failure with values deduced from fractographic analysis. This analysis helps to validate the use of fracture mechanics approaches for assessing damage tolerance in gas turbine engine components subjected to combined steady and vibratory stresses.Copyright © 1995 by ASME

Maximum likelihood is used to develop a quantitative means for defining confidence bounds on single crack growth curves and for comparing more than one crack growth curve. The concept of likelihood is discussed, the technique is applied to the Paris crack growth law, and statistical tests appropriate for pertinent comparative situations are reviewed. Examples of each test are examined, contrasted, and compared, and a practice for engineering comparisons of crack growth rate curves is suggested. Examples using data generated in the Fracture Mechanics Lab at Pratt & Whitney are presented and discussed. These discussions will lead to a final point regarding the role of statistical methods in augmenting engineering judgment and experience.

Susan E. Cunningham

Papers

Damage Tolerance Based Life Prediction in Gas Turbine Engine Blades Under Vibratory High Cycle Fatigue

Damage Tolerance Based Life Prediction in Gas Turbine Engine Blades Under Vibratory High Cycle Fatigue

Using Maximum Likelihood Techniques in Evaluating Fatigue Crack Growth Curves