Power Generation Degradation Modeling with Adaptive Transfer Functions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing power generation system control methods face challenges in accurately quantifying degradation due to factors like machine-to-machine manufacturing variations, errors in base models, and uncertainties in inputs, which complicates the adjustment of correction factors and affects the accuracy of physics-based software models.

Innovation Solution

A controller system that uses a processor to generate a model of the power generation system, adjusts correction factors based on live measurements, and represents these adjustments through transfer functions to quantify degradation, filtering out other sources of variation and improving model accuracy over time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If comparison of model outputs over time is used to quantify degradation, then degradation can be tracked, but accuracy is reduced due to influence from manufacturing variation, model errors, and input uncertainties

Engineering Contradiction:
Improvedegradation quantificationVSAvoiddegradation measurement accuracy
Core Design Contradiction:
ReliabilityVSMeasurement precision

Solution Approach 1:

The patent segments the correction factor into multiple independent components: a time-varying degradation component and static components representing manufacturing variation and model errors. By separating these effects, the degradation signal can be extracted without contamination from other sources of variation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system uses feedback from live measurements to continuously update and refine the degradation model. The measured parameters are compared with model predictions, and the difference (residual) is used to update the degradation state, creating a closed-loop system that improves accuracy over time.

Inventive Principle:
Principle #23Feedback

2Measurement precision

If correction factors are adjusted to match live measurements to model outputs, then model accuracy improves, but it becomes difficult to distinguish degradation from other sources of error

Engineering Contradiction:
Improvemodel accuracyVSAvoiddegradation signal isolation
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent performs preliminary characterization of the static components (manufacturing variation and model errors) before degradation analysis. By establishing baseline values for these static factors in advance, the system can subtract their influence from the correction factor, leaving only the degradation component to be analyzed.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system applies correction factors selectively - using full correction for model accuracy while using a separated, time-varying portion for degradation tracking. This partial application of correction allows the system to maintain model accuracy while preserving degradation information that would otherwise be masked.

Inventive Principle:
Principle #16Partial or excessive action

3Adaptability or versatility

If machine-to-machine manufacturing variation is accounted for in the model, then model applicability across units improves, but degradation detection becomes more complex

Engineering Contradiction:
Improvemodel applicabilityVSAvoiddegradation detection complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies local quality by allowing each power generation unit to have its own static correction factors that capture unit-specific manufacturing variations. Meanwhile, the degradation component follows a universal time-varying pattern that can be tracked across all units using the same methodology, simplifying the overall detection process.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3255512B1Adaptive model-based method to quantify degradation of a power generation system
Publication Date: 2021.10.20 GENERAL ELECTRIC CO
  • EP3255512B1 patent drawingFigure 1
  • EP3255512B1 patent drawingFigure 2
  • EP3255512B1 patent drawingFigure 3

AI summary

A system includes a power generation system (10) and a controller (18) that controls the power generation system (10). The controller (18) includes a processor that generates a model (40) of the power generation system (10) that estimates a value (44,48) for a first parameter (44,46) of the power generation system (10). The processor also receives a measured value of the first parameter (44,46). The processor further adjusts a correction factor (52) of the model (40) such that the estimated value (44,48) of the first parameter (44,46) output by the model (40) is approximately equal to the measured value of the first parameter (44,46). The processor also generates a transfer function (86,88,90,126,128,130) that represents the correction factor (52) as a function of a second parameter (42) of the power generation system (10). The processor further displays the transfer function (86,88,90,126,128,130) along with one or more previously generated transfer functions (86,88,90,126,128,130) to quantify degradation of the power generation system (10).