Monitoring Target Selection for Multi-Cycle Anomaly Diagnosis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing abnormality detection systems in nuclear power generation plants face difficulties in accurately detecting anomalies due to discontinuous changes in measurement information from devices with different operation cycles, leading to inefficient and erroneous detection.
Innovation Solution
A monitoring target selecting device that classifies measurement parameters across different operation cycles, selects relevant parameters based on mean values and inclinations, and outputs them to an abnormality sign monitoring device to reduce erroneous detection and improve anomaly identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If measurement information from devices with different operation cycles is used for abnormality detection, then the monitoring coverage is improved, but the detection accuracy deteriorates due to correlation breakdown
Solution Approach 1:
The patent segments the measurement parameters into different groups based on their operation cycles. By classifying parameters according to their cyclic characteristics, the system can apply appropriate correlation analysis methods to each group, thereby maintaining detection accuracy while covering multiple operation cycles.
Solution Approach 2:
The patent performs preliminary classification of measurement parameters before correlation analysis. By pre-organizing parameters according to their operation cycles and characteristics, the system prepares the data in advance to avoid correlation breakdown during the detection process, thus maintaining accuracy across different cycles.
2Reliability
If all measurement parameters are monitored continuously, then the detection completeness is improved, but the computational complexity increases
Solution Approach 1:
The patent extracts and selects only the relevant measurement parameters that are necessary for abnormality detection based on their operation cycles and correlations. By removing redundant parameters from continuous monitoring, the system reduces computational complexity while maintaining detection completeness for critical parameters.
Solution Approach 2:
The patent applies partial monitoring by focusing computational resources on parameters with higher abnormality indicators or those critical to safety. Instead of uniformly monitoring all parameters with equal intensity, the system adjusts monitoring depth based on parameter importance, reducing overall computational burden while maintaining detection effectiveness.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
A monitoring target selecting device is a monitoring target selecting device which is configured to output a measurement parameter to an abnormality diagnosis device that is configured to diagnose an abnormal event of a plant based on a correlation value representing a mutual correlation between measurement parameters output from the monitoring target selecting device, and includes a classification unit configured to acquire a plurality of measurement parameters measured in the plant, classify a change behavior of measured value over a time for each of the plurality of measurement parameters in a first period, and classify a change behavior of a measured value over a time for each of the plurality of measurement parameters in each of the first period and a second period, and a selection unit configured to select the measurement parameter as a measurement parameter to be output to the abnormality diagnosis device on the basis of a result of comparing a behavior of the measurement parameter in the first period and a behavior of the measurement parameter in the second period.