Statistical Model Combustion Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current linear analysis methods provide unreliable and insufficiently accurate results for understanding the influence of variables on combustion processes in combustion chambers, leading to unstable flame fronts and damage due to oscillations and pressure fluctuations in gas turbines.
Innovation Solution
A trainable statistical model, such as a neural network or Bayesian network, is used to analyze the influence of variables on combustion processes, allowing for both qualitative and quantitative assessments of their effects, and identifying cause-effect relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If linear analysis methods are used to investigate dependences between influencing variables and combustion process, then the analysis method is simple and easy to implement, but the results are unreliable and insufficiently accurate
Solution Approach 1:
The patent changes the fundamental parameter of the analysis method from linear to non-linear, using artificial neural networks that can capture complex non-linear relationships between combustion variables. This allows accurate modeling of the non-linear combustion process while maintaining practical applicability through trained models.
Solution Approach 2:
The patent replaces traditional linear statistical methods with a computational approach using artificial neural networks. This substitution enables the system to handle non-linear dependencies that linear methods cannot capture, significantly improving analysis accuracy for combustion processes.
2Reliability
If active instability control based on anticyclical modulation of fuel flow is implemented, then combustion oscillations can be eliminated, but costly sensor and actuator technology is required
Solution Approach 1:
The patent introduces a software-based intermediary (artificial neural network model) that processes combustion data and provides insights without requiring complex physical sensors or actuators. This intermediary enables analysis and prediction of combustion instability through computational methods rather than expensive hardware modifications.
Solution Approach 2:
The patent substitutes complex mechanical sensor and actuator systems with a computational software solution. The artificial neural network provides stability analysis and prediction capabilities without requiring the costly hardware infrastructure needed for active instability control systems.
3Object-affected harmful factors
If load on gas turbine is reduced to reduce combustion chamber humming, then damaging oscillations are reduced or avoided, but performance pledges to customers may not be fulfilled
Solution Approach 1:
The patent implements a feedback-based approach where the artificial neural network continuously analyzes combustion process data to identify variables influencing humming. This feedback mechanism enables real-time detection and analysis of combustion instability causes, allowing for targeted interventions that maintain power output while reducing harmful oscillations.
Solution Approach 2:
The patent enables analysis of how changes in operating parameters affect combustion chamber humming through the trained neural network model. This allows identification of optimal parameter combinations that maintain high power output while minimizing damaging oscillations, avoiding the need to simply reduce load.
4Reliability
If detailed investigation of oscillation phenomenon and comparison of active and passive methods is conducted, then possibilities for removal of oscillations are identified, but the approach is time-consuming and complex
Solution Approach 1:
The patent performs preliminary action by training the artificial neural network model in advance with comprehensive combustion data. This pre-trained model can then quickly analyze new scenarios and provide insights without requiring time-consuming detailed investigations each time, significantly reducing analysis time while maintaining deep understanding of the combustion process.
Data Source
AI summary
The invention relates to sensitivity analysis of variables influencing a combustion process. A trainable, statistical model is trained in such a way that it describes the combustion process in the combustion chamber. The trained statistical model is used to determine the influence of the variables on said combustion process in the combustion chamber.