Sensor Relationship Fluctuation Monitoring for Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In control subjects with multiple sensors, such as plants, monitoring for anomalies is challenging due to the complexity of sensor interactions, requiring a more effective method to identify and judge anomaly occurrences.
Innovation Solution
A monitoring device and method that calculates the relationship fluctuation level between signal outputs of subject sensors and related sensors, identifies sensor combinations with significant fluctuations, and generates anomaly sign information, including moving average trends, to facilitate more accurate anomaly detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple sensors are provided in a control subject to monitor various sensing subjects, then the monitoring coverage and detection capability are improved, but the complexity of monitoring and judging anomaly occurrences increases
Solution Approach 1:
The patent combines multiple sensor signals into a unified anomaly judgment system. The anomaly judgment unit integrates signals from multiple sensors (flow rate sensor, temperature sensor, pressure sensor) and applies synthesis rules to determine anomalies, merging individual sensor monitoring into a coordinated system that reduces overall monitoring complexity while maintaining comprehensive coverage
Solution Approach 2:
The anomaly judgment unit serves multiple sensing subjects (piping, pumps, driving devices, heat generating devices) with a single monitoring system. This multi-functional unit handles anomaly detection across different sensor types and monitoring targets, reducing the need for separate monitoring systems for each sensing subject
2Measurement precision
If multiple sensors are monitored individually to detect anomalies, then the detection precision for each sensor is maintained, but the time required for monitoring and judgment increases
Solution Approach 1:
The system merges anomaly judgment operations by processing multiple sensor signals simultaneously through the anomaly judgment unit. Instead of sequentially analyzing each sensor, the unit synthesizes signals from multiple sensors in parallel and applies judgment rules to determine anomalies, reducing total monitoring time while maintaining detection precision
Solution Approach 2:
The anomaly judgment rules are predetermined and stored in advance. When sensor signals are received, the system applies pre-established synthesis rules and determination criteria immediately, eliminating the need for complex real-time analysis and reducing judgment time while maintaining accurate anomaly detection
Data Source
AI summary
Among a plurality of sensors that detect a state of a control subject, a relationship fluctuation level between a magnitude of a signal output by a subject sensor and a magnitude of a signal output by another related sensor is calculated for each subject sensor and related sensor combination. Moreover, the subject sensor and related sensor combinations for which a value of the relationship fluctuation level is a predetermined value or more are identified, and a respective number of appearances of the subject sensor and the related sensor included in the identified combinations thereof is calculated.