Abstract

The emergence of blade tip timing (BTT) data simulators, providing timing samples from their known blade tip vibration output, has been primarily driven by the necessity of developing and evaluating BTT signal processing algorithms, avoiding costly experimental set-ups. Apart from the realism required of its mathematical model, the simulator is required to produce data from all blades, and various configurations of probes, in feasible time frames, including the effect of axial deformation on the timing data. It should also have the capability to produce data with errors comparable to speed-dependent noise in real data. The novel contribution of this paper is the presentation of three novel bladed disk simulators that address these requirements, and their validation in a novel type of study that considers all blades of an aero-engine compressor, and which uses a commercial BTT software package to process the data and identify an excited blade mode-family member. Simulated data with speed-dependent errors amounting to 10% noise-to-signal ratio are successfully processed, resulting in mean errors of 2.4%, 0.15% in resonant amplitude and frequency across the blades. The simulated BTT data are shown to correctly distinguish between two excited mode-pair members differing by just 1% in natural frequency. Correlation with experimental BTT data is also presented.