No abstract

No abstract

Blade vibrations, with the possibility of failure, is one of the major factors controlling the reliability of compressors and turbines. The prospects of encountering high alternating stress environments in blades make efficient turbomachine operation a very challenging task. In many cases the compressor or turbine functions through a wide range of load, flow, temperature, and speed which affect blade vibration, thus the stress environment continuously changes as the operating conditions changes. Any flow disturbance upstream of the rotating blades and some disturbances downstream will produce repetitive wake pulses that excite the blades. Resonance occurs with any coincidence of repetitive pulses with structural natural frequencies of rotating blades or impellers resulting in substantial amplification of alternating stresses. Most OEM design practices control vibratory stresses by avoiding resonance with expected stator sources; those excitations that cannot be avoided are designed with sufficient endurance to prevent failure. Thus three aspects of rotor/ blade design affect reliability: 1) aerodynamic excitation level and frequency, 2) structural response and resonance margins, and 3) selection and control of materials, coatings and their fabrication process to withstand the service environment. The main objective of this study is to develop a mathematical model to simulate the stresses in the rotating blade row that evaluates all three aspects of design to assess long term endurance. This is a two part paper on high cycle fatigue (HCF) failure analysis procedure of rotating blades and impellers. Part 1 [1] discusses aerodynamic excitation caused by stator vane and its role in generation of blade vibration. Here comprehensive computational fluid dynamics (CFD) is used to get a better understanding of the stator-rotor flow interactions at different operating conditions. The results of the aerodynamic simulations are order related excitation spectrum that can be applied to the stress/pulsation relationship defined in this part of the paper. This paper, Part 2, discusses an empirical dynamic stress model developed by impulse testing, assessing material endurance strength, and evaluation of criteria for failure by HCF.

Stress levels predicted by conventional disk modeling assumptions are lower than expected to cause conventional creep or fatigue damage consistent with slot failures experienced in some compressor and turbine disks. It was suspected that disparate slot to slot friction at the blade root surface will result in sticking of some blade roots as the turbine is shut down while adjacent blades slip; the un-resisted stuck root would pry the steeples apart causing additional bending stress. Testing of a blade root/disk slot pair in a load frame found that the blade root will stick in place as imposed radial loads decrease. Simulation of blade root movement during shutdown indicates peak stress can increase by 20% or more depending on geometric factors. The slot stress only rises above its maximum speed condition on shutdown (at 80% Max Speed in the example case). This brief stress rise will not cause significant creep damage, but can shorten disk life based on low cycle fatigue or hold time fatigue damage.

A multi-objective technique for designing single stage compressors is described. The blade geometry is defined as a function of known parametric equations, and the coefficient multipliers are viewed as design variables. A design of experiment approach is used to reduce the large combinations of design variables into a smaller subset. A response surface method is used to approximate the performance (total pressure rise, and adiabatic efficiency) as a function of design variables, and an optimum configuration is determined. This method has been applied to a rotor-stator combination similar to NASA Stage 35 with encouraging results.