Advanced Multi-Source Digital Signal Processing And Analysis For Characterization Of Vibrations In Turbo-Machinery Blades

Characterizing the dynamic response of advanced turbomachinery designs is often challenged by the real-world limits in the amount of instrumentation that is available, and by instrumentation longevity. By combining data from multiple sources, and tightly integrating the test results with physics-based models, this project will provide a unique data processing capability that has been validated for characterization of complex vibratory response of turbomachinery blades. Although great strides have been made, the accuracy of the NSMS data processing can suffer from the built-in assumptions regarding the vibratory mode shapes and the resulting stress distributions. This project addresses those limitations by integrating test results with physical models to provide the following capabilities: • Include virtual sensors in the pre- and post-processing of CFD, FEA, mistuning, and rotordynamic models, allowing the user to generate analytical predictions that correspond to actual measurements from the engine. • Include automatic functions that fine-tune the models to reconcile the predicted response with the measured data in terms of stochastic objective functions (accuracy and uncertainty). • Implement hybrid data reduction algorithms that use NSMS data in conjunction with other data sources to improve the understanding of complex dynamic behavior.