Sensor Fault Detection Using Confidence Ratings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Sensor faults in asset health management systems compromise the accuracy of diagnostic and prognostic information, making it difficult to identify faulty sensors, leading to unnecessary asset replacement, downtime, and increased costs due to the variety of sensors used.
Innovation Solution
A method and system that identify the appropriate tools for data acquisition systems based on characteristics to determine a confidence rating, utilizing tools like limit detection, slope calculation, autocorrelation, and cross-correlation to assess sensor reliability and operational status, thereby detecting sensor faults effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple different types of sensors are used in the asset health management system, then the system can monitor a wider range of asset conditions, but it becomes very difficult to determine when a sensor may have a fault
Solution Approach 1:
The patent applies universality by creating a unified fault detection framework that works across multiple sensor types (vibration, temperature, pressure, flow, etc.). The system uses a common set of analytical tools including autocorrelation analysis, spectral analysis, and trend analysis that can be applied universally to different sensor data, eliminating the need for type-specific detection methods while maintaining effectiveness across diverse sensor technologies.
Solution Approach 2:
The patent employs parameter changes by transforming sensor data into different analytical domains (time domain, frequency domain, statistical domain) to detect faults. By changing the parameters of analysis rather than the sensors themselves, the system can detect faults across various sensor types using consistent mathematical transformations and statistical measures, making fault detection independent of sensor type.
2Productivity
If sensor faults are not detected, then the system continues operating, but the accuracy of diagnostic and prognostic information is substantially compromised
Solution Approach 1:
The patent implements feedback by continuously monitoring sensor data quality metrics and using this information to adjust diagnostic interpretations. The system provides feedback loops where sensor readings are constantly evaluated against expected patterns, and when anomalies are detected, the system adjusts its diagnostic confidence levels and can trigger alerts for sensor validation, ensuring that diagnostic accuracy is maintained through continuous quality assessment.
Solution Approach 2:
The patent applies preliminary action by establishing baseline sensor performance characteristics and expected data patterns before faults occur. The system pre-configures detection thresholds, normal operating ranges, and correlation expectations for different sensor types, enabling it to quickly identify deviations that indicate faults. This preliminary preparation allows the system to maintain diagnostic accuracy by having reference standards ready for comparison.
Data Source
AI summary
A method, computer readable medium and system for detecting a sensor fault includes identifying one or more of a plurality of tools to use with at least one of a plurality of data acquisition systems based on data obtained from and at least one characteristic of the at least one of the data acquisition systems. The identified one or more tools are utilized on the obtained data to determine at least one confidence rating. An operational status for the at least one of the data acquisition systems is determined and provided based on at least the one determined confidence rating.