Sensor Health Monitoring for Duplicate Anomaly Alert Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Industrial asset monitoring systems often misdiagnose asset health due to unaccounted sensor health issues, leading to duplicate alerts and user bandwidth exhaustion, as they fail to differentiate between related anomalies and correctly assess data quality.
Innovation Solution
Implementing machine learning techniques for sensor health monitoring, including data quality verification and anomaly relation identification, to suppress duplicate alerts and improve predictive maintenance by estimating missing data and assessing its quality, thereby reducing misdiagnosis and alert fatigue.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rule-based detection schemes are used to monitor equipment performance, then anomaly detection capability is provided, but misdiagnosis occurs due to unaccounted sensor health issues and duplicate alerts are generated
Solution Approach 1:
A data quality verification layer is introduced as an intermediary between sensor data collection and anomaly detection. This layer assesses data quality metrics (completeness, consistency, plausibility) and filters or flags low-quality data before it reaches the rule-based detection schemes, preventing misdiagnosis caused by poor data quality while maintaining the existing anomaly detection capability
Solution Approach 2:
Data quality verification is performed preliminarily before anomaly detection analysis. By pre-assessing data quality and removing or flagging poor-quality data samples, the system prepares clean input data for the detection schemes, eliminating the root cause of misdiagnosis and duplicate alerts before they occur
2Productivity
If sensor data is continuously monitored without data quality verification, then monitoring coverage is maintained, but user bandwidth is exhausted due to duplicate alerts
Solution Approach 1:
Instead of verifying every single data point, the system applies data quality verification selectively to data samples that are likely to trigger anomalies or show quality degradation trends. This partial verification approach maintains monitoring coverage while reducing the number of false alerts that consume user bandwidth
Solution Approach 2:
The system implements feedback loops where alert patterns and data quality metrics are continuously monitored. When duplicate alerts or quality issues are detected, the system adjusts verification intensity and thresholds dynamically, reducing unnecessary alerts while maintaining coverage for critical conditions
Data Source
AI summary
Systems, methods, and computer readable storage mediums for performing sensor health monitoring are described. The method includes verifying data quality and suppressing alert generation using machine learning techniques to identify whether two anomalies generated by an asset monitoring system are related. The method can include receiving data characterizing measurement data acquired by a sensor coupled to an industrial asset. An anomalous data sample within the received data can be identified and removed from the anomalous data sample. A new sample of the removed data sample can be estimated using interpolation and the new sample can be assessed. Maintenance analysis can be performed based on the assessed, estimated new sample.