Blade Tip Timing Analysis for Gas Turbine Resonance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for analyzing blade tip timing (BTT) data in gas turbine engines face challenges such as undersampling, noise, and the need for separate synchronous and asynchronous analysis, making online analysis complex and requiring determination of response types.
Innovation Solution
A method that identifies resonant vibration events by zeroing blade displacements, fitting modelled displacements to multiple frequencies, and characterizing events by correlating frequencies with zeroed displacements, allowing for automated and online analysis without distinguishing between synchronous and asynchronous responses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If blade tip timing data is collected using stationary probes, then the complexity of the telemetry system is reduced, but the sampling rate falls below the Nyquist frequency causing undersampling and aliasing
Solution Approach 1:
The patent applies dynamics by making the analysis method adaptive to varying rotational speeds. The system dynamically adjusts the analysis approach based on whether the sampling rate is above or below the Nyquist frequency, allowing accurate vibration characterization despite the fixed low sampling rate of stationary probes.
Solution Approach 2:
The patent changes the analysis parameters based on the sampling conditions. When undersampling is detected, the system switches to a different analysis methodology that accounts for the sub-Nyquist sampling rate, effectively transforming the measurement approach to match the measurement constraints.
2Measurement precision
If separate synchronous and asynchronous analysis methods are used, then the precision of vibration characterization is improved, but the complexity of the analysis process increases
Solution Approach 1:
The patent merges synchronous and asynchronous analysis methods into a single unified computational framework. The system evaluates multiple potential vibration frequencies simultaneously and automatically identifies the dominant frequency, eliminating the need for separate analysis procedures while maintaining measurement precision.
Solution Approach 2:
The patent creates a universal analysis method that handles both synchronous and asynchronous vibrations through a single process. The methodology is multi-functional, capable of detecting and characterizing different types of vibration events without requiring the user to determine the response type in advance.
3Adaptability or versatility
If the rotational speed varies, then the adaptability of the engine operation is improved, but the determination of synchronous versus asynchronous response becomes difficult
Solution Approach 1:
The patent applies dynamics by making the frequency identification process adaptive to varying rotational speeds. The system continuously monitors the correlation between predicted and actual blade timings across multiple rotations, dynamically identifying the dominant vibration frequency regardless of whether the engine operates synchronously or asynchronously.
Solution Approach 2:
The system performs self-service by automatically determining the dominant vibration frequency without requiring external input about the engine's operational mode. The algorithm independently analyzes the timing data and identifies resonant events, eliminating the need for operators to classify responses as synchronous or asynchronous.
Data Source
Figure 1~2
Figure 3~4a
Figure 4b
AI summary
A method of analysing blade displacements is provided. The blade displacements are detected by a plurality of circumferentially spaced stationary timing probes associated with an assembly of rotating blades mounted on a rotor. The blade displacements correspond to the times at which the blades pass the respective probes. The method including the steps of: (a) identifying a possible resonant vibration event in the assembly of rotating blades; (b) zeroing the blade displacements on the rotations identified with the resonant vibration event to remove invariant blade displacements; (c) fitting modelled blade displacements corresponding to possible blade vibrational deflections at a plurality of frequencies to the zeroed blade displacements; and (d) characterising the resonant vibration event by identifying at each rotation the frequency having modelled blade displacements which correlate best with the zeroed blade displacements. Step (c) includes performing at each individual rotation identified with the resonant vibration event the sub-step of: (c-i) fitting the modelled blade displacements at each frequency to the zeroed blade displacements for a plurality of successive rotations which include that individual rotation.