Time-Series Event Extraction for Equipment Abnormality Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing time-series data processing technologies cannot effectively detect equipment abnormalities without event information, as they rely on pre-identified events, which may not be available or easily accessible, especially when recorded on non-electronic media.
Innovation Solution
A time-series data processing device that generates event information by extracting event waveforms from time-series data, calculates co-occurrence rates, and determines event conditions using a co-occurrence rate calculating unit and event information generating unit, allowing for abnormality detection even without pre-provided event information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If event information is provided externally, then abnormality detection accuracy is improved, but system complexity and data acquisition difficulty increase
Solution Approach 1:
The system extracts event information automatically from time-series data itself without requiring external event information sources. The event information generation unit analyzes patterns in the time-series data to identify events, making the system self-sufficient and eliminating the need for complex external data acquisition systems.
Solution Approach 2:
The patent introduces an event information generation unit as an intermediary that bridges the gap between raw time-series data and abnormality detection algorithms. This unit processes time-series data to generate structured event information, which then feeds into the abnormality detection process, simplifying the overall system architecture.
2Loss of information
If event information is recorded on print medium, then information availability is improved, but electronic processing capability deteriorates
Solution Approach 1:
The patent replaces the mechanical system of print medium with an electronic data storage and processing system. Time-series data is stored electronically and processed by the event information generation unit, eliminating the need for physical print media while preserving information availability and enabling automated electronic processing.
3Measurement precision
If outlier detection is performed on time-series data, then anomaly identification is improved, but false positive rate increases
Solution Approach 1:
The patent segments the time-series data into meaningful events using the event information generation unit. By dividing the continuous data stream into discrete events with specific characteristics, the system can analyze each event contextually, reducing false positives while maintaining anomaly detection sensitivity.
Solution Approach 2:
The system uses generated event information as feedback to refine abnormality detection. The event information provides contextual understanding that helps distinguish between normal variations and actual anomalies, reducing false positives while maintaining high detection accuracy.
Data Source
Figure 1
Figure 2A~2B
Figure 3
AI summary
An event waveform extracting unit (3) extracts an event waveform from time-series data. A co-occurrence rate calculating unit (4) calculates co-occurrence rates of event waveforms among the time-series data. A grouping unit (5) classifies the time-series data into groups depending the co-occurrence rates of the event waveforms. An event information generating unit (6) determines the time at which the periods during which event waveforms occur overlap with each other among the time-series data included in each group, and generates event information identifying an event related to the event waveforms on the basis of the determined time.