Manufacturing Anomaly Detection Using Cycle-Time Variable Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Complex industrial manufacturing processes with numerous IIoT devices make it challenging to detect anomalies and malfunctions, affecting overall equipment effectiveness (OEE) and quality assurance (QA).
Innovation Solution
A computer-implemented method that involves obtaining historic process variables, determining cycle times, clustering variables by cycle time, arranging clusters into a hierarchical tree, and detecting anomalies by comparing current cycle times to historic values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a hierarchical tree structure is used to organize process variables, then complexity reduction is achieved, but the time and computational resources required for clustering and tree construction increase
Solution Approach 1:
The patent performs clustering and hierarchical tree construction as preliminary actions during a setup phase or offline period, before actual anomaly detection begins. This allows the complex computational work to be done in advance when it doesn't interfere with real-time monitoring, thus reducing operational complexity without adding time loss during critical detection phases
Solution Approach 2:
The patent segments the large set of process variables into smaller clusters based on cycle time similarities, and further organizes these clusters into a hierarchical tree structure. This segmentation breaks down the complex task of analyzing all variables simultaneously into manageable groups, reducing overall system complexity while enabling efficient anomaly detection
2Measurement precision
If cycle time-based clustering is applied to process variables, then anomaly detection precision is improved, but the computational complexity of determining and comparing cycle times increases
Solution Approach 1:
The patent changes the parameter used for clustering from raw process variable values to cycle time characteristics. By transforming the data representation to focus on temporal patterns rather than absolute values, the system achieves better anomaly detection precision while the computational complexity is managed through efficient cycle time calculation algorithms
3Reliability
If all process variables from numerous IIoT devices are monitored, then comprehensive anomaly detection is achieved, but the difficulty of identifying relevant variables and detecting anomalies increases
Solution Approach 1:
The patent segments the comprehensive set of process variables from multiple IIoT devices into clusters based on their cycle time characteristics. This segmentation maintains comprehensive monitoring coverage while organizing variables into manageable groups, making it easier to identify relevant variables for anomaly detection without losing reliability
Solution Approach 2:
The patent introduces a new dimension for organizing variables by adding the hierarchical tree structure based on cycle time relationships. This dimensional organization transforms the flat, overwhelming list of variables into a structured hierarchy, making relevant variables easier to identify while maintaining comprehensive anomaly detection capability
Data Source
AI summary
A method for detecting an anomaly includes obtaining a time-series of historical process variables within a predefined time span; determining a cycle time of the historical process variables; clustering the historical process variables into clusters based on cycle time; arranging the clusters into a tree; storing the tree; obtaining a time-series of a plurality of current process variables, which correspond to the historic process variables; and detecting the anomaly of at least one device by identifying a cycle time of a current process variable that is longer than the cycle time of a corresponding historic variable.