Generator Flashover Prediction via Statistical Signal Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting collector flashovers in generators are inefficient, as they require frequent inspections and cannot predict potential flashovers during operation, leading to unexpected outages.
Innovation Solution
A system and method that uses a data acquisition system to receive and analyze measured variables, fitting a statistical model to estimate parameters like field circuit impedance and noise, predicting flashovers based on these estimates to prevent operational disruptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If periodic inspections are carried out to detect collector flashover, then the reliability of generator operation is improved, but the loss of time and productivity increases due to frequent shutdowns
Solution Approach 1:
The patent replaces mechanical inspection methods with signal processing and statistical analysis of electrical parameters. By analyzing field current and voltage signals through spectral analysis and statistical modeling, the system detects flashover conditions without requiring physical inspection or shutdown, thus maintaining reliability while eliminating time loss from frequent generator stoppages
Solution Approach 2:
The patent introduces an intermediary detection system that monitors electrical signals (field current, voltage) as intermediate parameters between normal operation and flashover failure. This intermediary monitoring enables early warning of flashover conditions without requiring direct intervention or shutdown, resolving the contradiction between maintaining reliability and avoiding time loss
2Reliability
If frequent inspections are performed to prevent collector flashover, then the reliability is improved, but the ease of operation deteriorates due to operational disruptions
Solution Approach 1:
The patent enables the generator monitoring system to perform self-diagnosis by continuously analyzing its own electrical signals for signs of flashover. This self-monitoring capability maintains high reliability through continuous assessment while preserving ease of operation, as the system automatically detects issues without requiring operator intervention or shutdown for inspection
3Device complexity
If existing detection methods are used, then the device complexity is kept low, but the measurement precision deteriorates as potential flashovers cannot be predicted during operation
Solution Approach 1:
The patent performs preliminary analysis of electrical signals to detect early signs of flashover before actual failure occurs. By continuously analyzing field current and voltage signals for statistical deviations and spectral changes, the system achieves precise prediction of potential flashovers during operation while maintaining relatively simple implementation through standard signal processing techniques
Data Source
AI summary
A system, method and computer program product for predicting collector flashover is disclosed. The method includes receiving a frame of measured data from a data acquisition system, fitting a statistical model to the measured data wherein the measured data includes a plurality of measured variables measured at multiple time instances. The statistical model and the measured data are used for estimating one or more parameters for the frame wherein the one or more parameters include at least one of field circuit impedance and field current noise. A flashover is predicted based, at least in part, upon the one or more estimated parameter for the frame.