Neural Network Steam Quality Estimation for OTSGs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for real-time steam quality measurement in Once-Through Steam Generators (OTSGs) are inadequate for precise control, as they rely on linear calculations with restrictive assumptions, are sensitive to measurement noise, and require frequent manual adjustments, failing to provide robust and timely feedback for optimal performance.
Innovation Solution
A system that uses sensors to obtain raw measurement values, determines a robustness index, and employs multiple models based on measurement reliability to estimate steam quality, with a corrector module filtering the estimates to reduce errors and adapt to changing conditions, allowing for real-time control of steam quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If linear function calculation is used for steam quality measurement, then the measurement can be performed continuously, but the measurement precision deteriorates due to restrictive assumptions and sensitivity to measurement noises
Solution Approach 1:
The patent changes the mathematical model from a simple linear function to a neural network-based non-linear model. This allows the system to capture complex relationships between process variables and steam quality without restrictive assumptions, thereby improving measurement precision while maintaining continuous operation capability.
Solution Approach 2:
The patent replaces traditional mechanical measurement systems with an intelligent sensing system that uses neural networks to process sensor data. This substitution enables more accurate steam quality estimation by leveraging pattern recognition and adaptive learning capabilities rather than relying on fixed linear relationships.
2Measurement precision
If manual sampling and lab analysis are used for steam quality measurement, then the measurement precision may be higher, but the productivity deteriorates due to manual procedures and lack of real-time feedback
Solution Approach 1:
The system performs self-measurement of steam quality using an integrated sensor array and neural network processor that continuously monitors and estimates steam quality without requiring external manual sampling or laboratory analysis. This self-service capability enables real-time feedback while maintaining high measurement precision through automated intelligent processing.
Solution Approach 2:
The patent implements a closed-loop feedback system where the neural network continuously processes sensor data to provide real-time steam quality estimates that feed back to the control system. This enables dynamic adjustment of operating parameters based on current steam quality conditions, achieving both high precision and real-time responsiveness.
3Reliability
If multiple models are used to estimate steam quality based on measurement reliability, then the robustness improves, but the device complexity increases
Solution Approach 1:
The patent implements a dynamic model selection mechanism where the neural network automatically switches between different estimation models based on the reliability and quality of input sensor measurements. This dynamic adaptation allows the system to maintain high robustness by selecting the most appropriate model for current operating conditions without requiring manual intervention or complex fixed architecture.
Data Source
AI summary
A system and method for estimating steam quality for a steam generator is provided which involves obtaining raw measurement values for process variables of the steam generator from sensors coupled to the steam generator; receiving the raw measurement values and slow-rate steam quality samples in order to determine a steam quality estimate, using a measurement module to receive the raw measurement values and determine model input values and a robustness index; using an estimator module to determines a raw steam quality estimate using the model input values and a model that is selected from several models depending on reliability of some of the raw measurements; and using a corrector module to determine the steam quality estimate using the raw steam quality estimate, robustness index, and slow-rate steam quality samples. An output receives the steam quality estimate and can provide the steam quality estimate to another device.