Multivariate Time-Series Anomaly Analysis for Error Element Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting abnormal data in processes using artificial neural networks struggle to identify specific elements causing errors, leading to difficulties in determining whether repair is needed for process or equipment elements.
Innovation Solution
A system and method that generate an anomaly detection model using multivariate time-series data, compare input data with reconstructed data, and calculate abnormality cause probabilities based on a database, providing detailed feedback on which process or equipment elements require maintenance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing detection models use artificial neural networks to detect abnormal data by comparing input data with output data, then abnormal data can be detected when the difference exceeds a certain level, but the specific elements causing errors cannot be identified
Solution Approach 1:
The patent segments the detection process into two distinct stages: first, detecting whether data is abnormal using a detection model; second, identifying specific abnormal elements using an analysis model that processes the detected abnormal data. This segmentation allows the system to maintain high detection accuracy while also providing detailed information about which specific elements are abnormal, thereby resolving the contradiction between detection precision and information completeness.
2Reliability
If existing detection models output data with lost information when abnormal data is input, then abnormal data can be detected based on difference threshold, but detailed feedback on specific process or equipment elements requiring repair cannot be provided
Solution Approach 1:
The patent implements a feedback mechanism where the analysis model takes the detected abnormal data as input and generates detailed feedback information identifying specific abnormal elements. This feedback loop transforms the incomplete output from the detection model into actionable information that clearly indicates which process or equipment elements require attention, thereby maintaining detection reliability while significantly improving ease of operation for maintenance decisions.
Data Source
AI summary
An operating method of a system for analyzing abnormal data includes generating an anomaly detection model by using a training data set including a plurality of pieces of multivariate time-series data, comparing first time-series data input to the anomaly detection model with second time-series data output from the anomaly detection model through an operation on the first time-series data, determining whether or not the first time-series data includes abnormal data, on the basis of the comparison between the first time-series data and the second time-series data, comparing a first plurality of data elements included in the first time-series data with a second plurality of data elements included in the second time-series data, and detecting at least one data element on the basis of a result of comparing the first plurality of data elements with the second plurality of data elements.