Multi-Can Combustor Dynamics Tuning via Sensor Data Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional combustion dynamics tuning algorithms for multi-can combustors face inefficiencies due to poor or errant sensor data, leading to decreased efficiency and potential damage from excessive vibration.
Innovation Solution
A method and system that utilize operating frequency information from multiple cans in a gas turbine engine to determine a median value, comparing it to thresholds to implement engine control actions and adjust frequencies, thereby mitigating the impact of poor sensor data and maintaining optimal combustion dynamics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional dynamics tuning algorithms use all sensor outputs including poor or errant data, then the algorithm can maintain simplicity in data processing, but combustion efficiency decreases and excessive vibration may occur
Solution Approach 1:
The patent extracts and removes poor or errant sensor data from the dataset before processing. The system identifies sensors providing poor measurements and excludes their data from the dynamics tuning algorithm, thereby preventing degradation of combustion efficiency while maintaining algorithm simplicity.
Solution Approach 2:
The patent introduces an intermediary data filtering step between sensor output and the dynamics tuning algorithm. This intermediary process evaluates sensor data quality and selectively passes only good measurements to the tuning algorithm, acting as a mediator that protects the system from poor data without complicating the core algorithm.
2Ease of operation
If conventional algorithms average all sensor outputs to determine dynamics signal, then processing remains simple, but tuning precision deteriorates due to poor sensor data
Solution Approach 1:
The patent extracts and removes poor sensor measurements from the averaging process. By identifying and excluding sensors providing errant data before computing the average dynamics signal, the system maintains computational simplicity while significantly improving the accuracy of the tuning signal.
Solution Approach 2:
The patent applies different quality assessments to different sensor data sources. Rather than uniformly averaging all sensor outputs, the system evaluates each sensor's data quality locally and applies selective weighting or exclusion, thereby improving overall measurement precision while keeping processing straightforward.
3Quantity of substance
If poor sensor data is input to the dynamics tuning algorithm, then data collection remains comprehensive, but combustion dynamics tuning effectiveness decreases
Solution Approach 1:
The patent extracts and removes poor quality sensor data from the comprehensive dataset before processing. By maintaining complete data collection for monitoring purposes while excluding errant measurements from the tuning algorithm input, the system preserves data comprehensiveness for diagnostics while ensuring tuning effectiveness.
Solution Approach 2:
The patent segments the sensor data into quality-based categories (good measurements versus poor/errant measurements). This segmentation allows the system to utilize all collected data for overall system monitoring while directing only high-quality data to the tuning algorithm, thereby maintaining both comprehensive data collection and effective tuning.
Data Source
AI summary
Embodiments of the invention can provide systems and methods for using a combustion dynamics tuning algorithm with a multi-can combustor. According to one embodiment of the invention, a method for controlling a gas turbine engine with an engine model can be implemented for an engine comprising multiple cans. The method can include obtaining operating frequency information associated with multiple cans of the engine. In addition, the method can include determining variation between operating frequency information of at least two cans. Furthermore, the method can include determining a median value based at least in part on the variation. Moreover, the method can include determining whether the median value exceeds at least one operating threshold. The method can also include implementing at least one engine control action to modify at least one of the operating frequencies if at least one operating threshold is exceeded.