Early Sign Detection for Sudden Vibration in Gas Turbines
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting sudden variant vibrations in machines, such as gas turbines, struggle to identify early signs effectively due to the rapid onset of these vibrations, making it difficult to prevent device damage.
Innovation Solution
An early sign detection device with multiple sensors measuring physical parameters at various positions, acquiring time series data, and computing occurrence probabilities of vibration modes based on amplitude and phase to detect sudden variant vibrations in advance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If control is executed after detection of vibration increase, then device damage can be prevented, but sudden variant vibration cannot be avoided because the vibration reaches limit cycle in short time
Solution Approach 1:
The system performs preliminary analysis of vibration data to compute occurrence probabilities of vibration modes before the actual sudden variant vibration occurs. By continuously monitoring and analyzing vibration characteristics in advance, the system can detect early signs and trigger control actions before the vibration reaches the limit cycle, thus preventing device damage while accounting for the short time available.
Solution Approach 2:
The system prepares control measures in advance by identifying patterns and computing occurrence probabilities of vibration modes that precede sudden variant vibrations. This allows the control system to be pre-positioned and ready to act immediately when early signs are detected, cushioning against the rapid progression to limit cycle vibration.
2Measurement precision
If network entropy analysis is performed on time series variation data from one position, then combustion vibration can be detected, but occurrence probability of vibration mode cannot be found and early sign detection is insufficient
Solution Approach 1:
The system segments the vibration analysis by identifying and analyzing different vibration modes separately. Instead of treating all vibration data uniformly, the system divides the time series data into components corresponding to different vibration modes and computes the occurrence probability of each mode independently, thereby recovering information that would be lost in aggregate analysis.
Solution Approach 2:
The system adds a new dimension to the analysis by computing occurrence probabilities as an additional parameter alongside traditional vibration detection metrics. This transforms the analysis from a single-dimensional detection approach to a multi-dimensional approach that includes probability assessment, enabling more comprehensive early sign detection.
Data Source
AI summary
An early sign detection device includes a plurality of sensors, a data acquisition unit configured to acquire time series variation data of a physical amount from each of the plurality of sensors, a computation unit configured to compute an occurrence probability of a vibration mode involved in a sudden variant vibration of a detection subject from an amplitude or a phase of the time series variation data of the physical amount at each of the plurality of positions, and a detection unit configured to detect an early sign of the sudden variant vibration on the basis of the occurrence probability of the vibration mode. The plurality of sensors are respectively disposed at a plurality of positions of the detection subject and are each configured to measure the physical amount at a corresponding position of the plurality of positions.