Combustor Blowout Precursor Detection Using Multifractal Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Combustion systems face challenges in predicting and preventing lean blowout, which can lead to unexpected shutdowns and safety issues due to the lack of robust methods for detecting blowout precursors in aircraft engines and gas turbines.
Innovation Solution
A system and method using Hurst exponent estimation, Burst count estimation, and recurrence quantification based estimation to analyze time series signals from dynamic state variables, enabling the detection of precursors to lean blowout and allowing for proactive control measures to prevent shutdowns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If spectral analysis, statistical analysis, or wavelet analysis is used to detect blowout precursors, then some prediction capability is achieved, but the prediction is not robust enough
Solution Approach 1:
The patent segments the blowout detection process into three distinct analytical components: spectral analysis for frequency domain characteristics, statistical analysis for temporal patterns, and wavelet analysis for time-frequency localization. Each segment addresses specific aspects of the precursor signals, and their combined results provide robust and precise prediction
Solution Approach 2:
The patent merges multiple analysis methods (spectral, statistical, and wavelet analyses) into a unified detection system. By combining the strengths of each method, the system achieves both robustness through multiple verification pathways and precision through complementary detection capabilities
2Object-generated harmful factors
If the combustor is operated in lean combustion mode to reduce NOX emissions, then emission levels are reduced, but the risk of complete flame loss increases
Solution Approach 1:
The patent implements preliminary detection of blowout precursors by continuously monitoring combustion parameters and analyzing their spectral, statistical, and wavelet characteristics. By detecting precursors before complete flame loss occurs, the system enables preventive action to maintain stable lean combustion while reducing NOX emissions
Solution Approach 2:
The patent establishes a feedback mechanism where detected precursors trigger control actions to adjust combustion parameters. This feedback loop allows the system to operate reliably in lean mode by continuously correcting deviations that could lead to flame loss, thus maintaining both low emissions and combustion stability
Data Source
AI summary
A method and system for determining one or more precursors to control blowout in a combustor is provided. The method includes obtaining a time series signal corresponding to a dynamic state variable of the combustor. The method includes detecting one or more precursor based on an analysis of the time series signal using one or more parameters to control blowout in the combustor. One or more parameters include a Hurst exponent estimation, a Burst count estimation, and a recurrence quantification based estimation.