Abnormality Detection Support Using Representative Sensor Time Series
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing system analysis methods require manual screening of sensor values and predictive models to accurately detect system abnormalities, which is inefficient and prone to errors, especially in large-scale systems where various events can affect abnormality scores.
Innovation Solution
A system analysis support device and method that uses a predictive model to calculate the overall abnormality degree of a system based on multiple time series data items, selecting and presenting time series data with similar changes to the abnormality degree, thereby automating the identification of relevant sensor data without manual screening.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual screening of sensor values and predictive models is performed to accurately detect system abnormalities, then detection accuracy is improved, but operation time and efficiency deteriorate
Solution Approach 1:
The system performs self-service by automatically selecting sensor values and predictive models that contribute significantly to abnormality detection. The determination unit autonomously identifies and screens sensor values based on their contribution to the abnormality degree, eliminating the need for manual screening while maintaining high detection accuracy.
Solution Approach 2:
The system changes the parameter of abnormality degree determination by introducing a contribution degree metric. Instead of manually evaluating sensor values, the system calculates the contribution degree of each sensor value to the overall abnormality degree, automatically identifying the most relevant sensors based on this quantitative parameter.
2Reliability
If all sensor values are analyzed to ensure comprehensive abnormality detection, then detection completeness is improved, but device complexity increases
Solution Approach 1:
The system extracts only the sensor values that have a significant contribution to the abnormality degree. The determination unit identifies and extracts relevant sensor values from the entire set of sensor data, focusing analysis on the most critical parameters while excluding less important ones, thus reducing system complexity without compromising detection completeness.
Solution Approach 2:
The system applies local quality by treating different sensor values differently based on their contribution degree. Instead of uniformly analyzing all sensor values, the system identifies and prioritizes specific sensor values that locally contribute most to abnormality detection, allocating computational resources efficiently to high-impact areas.
3Measurement precision
If multiple sensor values are processed to improve abnormality detection accuracy, then measurement precision is improved, but calculation amount increases
Solution Approach 1:
The system performs partial action by processing only the necessary subset of sensor values that contribute significantly to abnormality detection. Instead of calculating the contribution of all sensor values, the determination unit identifies and processes only those sensor values that exceed a certain contribution threshold, reducing the overall calculation amount while maintaining detection accuracy.
Data Source
AI summary
A system analysis support device includes a data acquisition part that obtains time series data (items) measured in a system that is to be analyzed, an overall abnormality degree calculation part that calculates transition of abnormality degree representing overall abnormality degree of the system that is to be analyzed, using a predictive model generated so that, with 2 or more time series data (items) as input, values representing a relationship between the 2 or more time series data (items) are outputted, and the time series data (items), and a representative index selection part that selects and presents time series data (items) indicating change similar to transition of the overall abnormality degree of the system that is to be analyzed, from among the time series data (items).