Kalman Filter Gas Turbine Modeling with Missing Data Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Gas turbine control systems face challenges in reconciling measured performance data when operating parameters such as net shaft power are unavailable or corrupt, leading to under-constrained or ineffective data reconciliation, especially in configurations where multiple components share a shaft.
Innovation Solution
The method involves generating predicted operating parameters, sensing actual parameters, determining differences, and using a Kalman filter to synthesize missing or corrupt data, applying adjustment parameters to the model to facilitate tuning and ensure accurate operation, even with incomplete data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a Kalman filter is used to reconcile measured gas turbine performance data, then the accuracy of performance monitoring is improved, but the system becomes under-constrained or ineffective when critical parameters like net shaft power are unavailable or corrupt
Solution Approach 1:
The patent modifies the Kalman filter configuration by changing the set of parameters used for reconciliation. When net shaft power is unavailable, the system switches to using alternative measurable parameters (compressor inlet pressure, compressor exit pressure, turbine exhaust temperature, generator power output) to constrain and solve the filter equations, maintaining effectiveness despite parameter changes in the input data
Solution Approach 2:
The patent creates a virtual representation of the missing or corrupt parameter through synthesis. By using the gas turbine cycle model and available measurements, the system generates a synthesized net shaft power value that replicates the information that would have been provided by direct measurement, allowing the Kalman filter to continue functioning with complete parameter sets
2Measurement precision
If net shaft power measurement is used as input to the Kalman filter, then data reconciliation accuracy is improved, but the system becomes difficult to implement in gas turbines with coupled steam and gas turbines where power contribution cannot be easily quantified
Solution Approach 1:
The patent extracts the net shaft power parameter from the set of required measurements by synthesizing it from other available parameters. Instead of requiring direct measurement of net shaft power (which is difficult in coupled configurations), the system extracts this information indirectly through model-based calculation using measurable parameters like generator power output and turbine exhaust temperature
Solution Approach 2:
The patent creates a universal Kalman filter configuration that can operate in multiple turbine configurations (single gas turbine, coupled gas-steam turbine, etc.). By designing the filter to work with a standardized set of measurable parameters and synthesizing net shaft power universally from these parameters, the system achieves ease of operation across different turbine architectures without requiring configuration-specific modifications
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method for modeling a gas turbine assembly (5) include a plurality of sensors (28) configured to measure a plurality of operating parameters (52) of the turbine assembly (5). The system further includes a model (38) configured to compare the plurality of operating parameters (52) with a plurality of predicted parameters (41) to determine difference values (62). A kalman filter (64) of system is configured to evaluate the difference values (62) to facilitate adjusting the model (38). The kalman filter (64) is further configured to accommodate for a missing operating parameter (52) of the plurality of operating parameters (52) to facilitate tuning the model (38).