ADI and ASU propose to develop a novel methodology for blade peak stress prediction from limited strain gage/tip-timing measurements when multiple vibratory modes are present. The current protocol assumes that only one mode is present and only provides upper and lower bound estimates of the blade peak stress when multiple modes are important. It is proposed here to construct the blade modal stresses distributions, not provided by the manufacturer, which lead directly to the blade peak stress estimation. Two approaches are proposed in Phase I to construct these distributions. They require only data provided by the manufacturer and basic blade geometry information, e.g. aspect ratio, taper, thickness, twist, stagger, referred to as “blade descriptors”. In the first approach (knowledge-based approach), a database of modal stresses distributions and natural frequencies is first created by finite element methods for a broad range of descriptor values. This database is then reduced and error estimates associated to it. In the second approach, a finite element model is constructed on the fly from blade descriptors and measured blade natural frequencies. Then, the modal stresses distributions are determined. Blind validation cases are planned to assess the reliability of the two approaches.
Benefit: The proposed methodology and software will permit an increased reliability of the predicted blade peak stress in experiments which will lead to an improved estimation of their high cycle fatigue life. In turn, improvements in life prediction will permit a longer running life (financial savings) and an increased accuracy in predicting failure (safety improvements) These improvements will be achieved without requiring undue characterization of the blade, i.e. only basic blade geometry information will be necessary, and thus will be usable by all testing facilities focusing on engine/turbomachine. Potential customers of the planned software will thus be Air Force, Navy, and Army but also NASA (NASA Glenn and NASA Marshall). Further, electric utilities, especially in conjunction with EPRI, are also potential users of the software. Note finally that the proposed effort also supports ongoing blade health monitoring programs.
Keywords: Stress estimation, Vibratory modes, bladed disks, Experimental vibration, peak stress, strain gages, tip-timing measurement, finite element method.
