Sensor Anomaly Detection With Dynamic Consecutive Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In large or complex infrastructure systems, such as power plants and manufacturing equipment, monitoring multiple sensors simultaneously is challenging with a limited number of personnel, and existing anomaly detection methods face a tradeoff between early anomaly detection and erroneous detection, where setting thresholds uniformly can lead to either early detection or reduced mis-detection but not both simultaneously.
Innovation Solution
An anomaly detection apparatus that includes a processing circuit to acquire sensor data, generate predicted values, set dynamic thresholds based on feature quantities, and count consecutive deviations to detect anomalies, thereby balancing early detection and mis-detection prevention by adjusting thresholds according to the frequency distribution of feature quantities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If thresholds are uniformly set to small values for the number of consecutive times, then anomalies can be detected early, but erroneous detection increases in number
Solution Approach 1:
The patent applies dynamics by making the threshold value changeable based on the temporal distribution characteristics of sensor data. Instead of using a fixed uniform threshold, the system dynamically adjusts the threshold according to whether the data exhibits temporal correlation, allowing early anomaly detection when appropriate while preventing erroneous detection in other scenarios
Solution Approach 2:
The patent changes the parameter of threshold values from static to dynamic based on feature quantities derived from temporal data distribution. By analyzing temporal correlations and adjusting thresholds accordingly, the system resolves the contradiction between early detection and reduced false positives
2Reliability
If thresholds are uniformly set to large values for the number of consecutive times, then erroneous detection can be decreased in number, but anomalies cannot be detected early
Solution Approach 1:
The system dynamically adjusts threshold values based on the temporal distribution characteristics of the data. When temporal correlation is detected, lower thresholds enable early anomaly detection; when no temporal correlation exists, higher thresholds prevent erroneous detection, thus resolving the contradiction
3Area of stationary object
If a limited number of people monitor all sensors simultaneously in large or complex systems, then monitoring coverage is maintained, but monitoring effectiveness decreases
Solution Approach 1:
The patent replaces manual mechanical monitoring with an automated information processing system that analyzes sensor data temporally. This substitution enables effective monitoring of large numbers of sensors without requiring proportional increases in human operators, maintaining coverage while improving effectiveness
Solution Approach 2:
The system introduces an intermediary automated detection mechanism that processes sensor data and identifies anomalies based on temporal patterns. This intermediary handles the complexity of monitoring numerous sensors, freeing human operators from direct monitoring tasks while maintaining comprehensive coverage
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
An anomaly detection apparatus includes a processing circuit. The processing circuit is configured to: acquire measured values from sensors installed in a system, a first function, a first threshold, and a second function to output a second threshold; generate the predicted values based on the measured value and the first function; detect that a deviation between the measured values and the predicted values exceeds the first threshold; calculate the feature quantities based on the measured values; and determine whether a number of consecutive times is equal to or larger than the second threshold to detect an anomaly or a sign of the anomaly.