Sensor Fall-Curve Analysis for Fault Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing IoT systems face challenges in accurately identifying sensor failures due to faulty data mimicking non-faulty data, and the inability to isolate the root cause of anomalies, especially in harsh environments where technical expertise may be limited.
Innovation Solution
A Fall-curve based technique is used to identify sensor faults by shutting off the sensor and analyzing the characteristic decay curve, which is unique to each sensor type and independent of the sensing environment, allowing for fault detection without additional hardware or spatiotemporal correlation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a data centric rule-based approach is used to detect sensor anomalies, then fault detection capability is provided, but faulty sensor data can mimic non-faulty data leading to inaccurate identification
Solution Approach 1:
Instead of analyzing sensor data during operation to detect faults, the patent applies inversion by turning off the sensor and analyzing the fall curve characteristics. This reverse approach allows identification of sensor-specific electrical characteristics that remain unique even when sensor data mimics normal readings, enabling accurate fault detection and root cause identification.
Solution Approach 2:
The patent changes the operational state parameter of the sensor from active sensing mode to powered-off mode. By analyzing the fall curve parameters (voltage decay characteristics) rather than sensor measurement data, the system can identify unique electrical signatures that distinguish faulty sensors from non-faulty ones, resolving the limitation where faulty data mimics normal data.
2Reliability
If existing rule-based schemes are used for fault detection, then anomalies can be detected, but the root cause of sensor failure cannot be isolated
Solution Approach 1:
The patent performs preliminary action by characterizing and storing the fall curve signatures of sensors during normal operation or initialization. This pre-established baseline allows subsequent comparison when anomalies occur, enabling not just detection but also identification of the specific faulty component by matching against known good sensor signatures.
Solution Approach 2:
The system implements feedback by continuously monitoring fall curve characteristics and comparing them against stored reference signatures. When deviations are detected, the feedback mechanism identifies the specific nature of the fault (e.g., open ADC connection, ground connection issue, sensor failure) by pattern matching, thereby recovering the lost root cause information.
3Ease of manufacture
If field staff with limited technical background are deployed for sensor maintenance, then operational costs are reduced, but the ability to identify and isolate sensor faults is diminished
Solution Approach 1:
The patent enables self-service by implementing automated fall curve analysis that performs diagnostic functions previously requiring expert technicians. The system automatically characterizes sensors, detects faults, and identifies root causes through fall curve pattern recognition, allowing field staff with limited technical background to perform comprehensive sensor diagnostics without specialized knowledge.
Data Source
AI summary
A computer implemented method includes turning off a sensor, receiving fall curve data from the sensor, and comparing the received fall curve data to a set of fall curve signatures to identify the sensor or a sensor fault.