Sensor Anomaly Detection via Upstream-Downstream Correlation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high cost and inefficiency of regular inspections and maintenance for large numbers of air and water quality sensors, particularly in widespread installations like offshore wind power areas, where random inspection methods often miss malfunctioning sensors.
Innovation Solution
A method of building upstream-and-downstream sensor configurations and anomaly detection using geographic location data, flow field data, and sensing data to determine correlations and identify satellite sensors, thereby reducing the need for extensive resource allocation and improving inspection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If random inspection method is used on sensors, then inspection operation is simple, but inspection efficiency is low and malfunction sensors may be missed
Solution Approach 1:
The system enables sensors to self-inspect by using their own sensing data and flow field data to determine their operational status. Each sensor autonomously compares its pollution detection data with flow field information to identify whether it is malfunctioning, eliminating the need for external random inspection operations.
Solution Approach 2:
The system implements a feedback mechanism where sensors continuously monitor their own performance by comparing their sensing data with expected patterns derived from flow field data. This feedback loop allows sensors to detect their own malfunctions and trigger appropriate responses, improving inspection efficiency without complicating operations.
2Reliability
If regular inspections are performed on large numbers of sensors, then sensor reliability is maintained, but time, labor and cost increase significantly
Solution Approach 1:
Sensors perform self-inspection by autonomously analyzing their own sensing data against flow field data patterns. This self-service approach maintains sensor reliability without requiring external inspection resources, thereby eliminating the time and labor costs associated with traditional regular inspections.
Solution Approach 2:
The system performs preliminary analysis by continuously monitoring and comparing sensing data with flow field data in real-time. This preliminary detection of potential malfunctions allows for early intervention, maintaining reliability while reducing the need for time-consuming periodic inspections.
3Measurement precision
If upstream-and-downstream relationship is established between sensors, then anomaly detection accuracy is improved, but system complexity increases
Solution Approach 1:
The flow field data serves as an intermediary element that connects upstream and downstream sensors. By using flow field information as a mediator, the system establishes relationships between sensors without requiring direct complex communication or control mechanisms, thus improving detection accuracy while limiting complexity increase.
Solution Approach 2:
The system segments the sensor network into upstream and downstream groups based on their spatial relationships and flow field characteristics. This segmentation allows for simplified anomaly detection by processing sensors in discrete groups rather than as a complex interconnected system, maintaining precision while managing complexity.
Data Source
AI summary
A method of building upstream-and-downstream configuration of sensors includes determining two sets of geographic position data of a target sensor and a candidate sensor, obtaining pollution-associated periods according to pieces of flow field data, the sets of geographic position data and pieces of target sensing data of the target sensor to determine a pollution-associated period, calculating a correlation between target sensing data obtained by the target sensor during the pollution-associated period and candidate sensing data obtained by the candidate sensor during the associated air pollution period to obtain sensor correlations, and determining the target sensor and the candidate sensor having a upstream-and-downstream relationship with the candidate sensor being used as a satellite sensor of the target sensor when a quantity ratio of sensor correlations being larger than or equal to a correlation threshold is larger than or equal to a default ratio.