Multi-dimensional Sequential Pattern Mining for Machine Fleet Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for monitoring machines, such as construction or hoisting machines, are limited in analyzing multi-dimensional sequential patterns across different but similar machine types, making it difficult to detect common patterns and provide effective preventive maintenance.
Innovation Solution
A method that transfers event data from machines to a central processor for mining multi-dimensional sequential patterns, where attributes indicate machine properties, and matches these patterns with a central database to identify correlations and automate maintenance actions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If single-dimensional event data mining is used, then the analysis is simpler and faster, but the ability to detect common patterns across machine fleets is limited
Solution Approach 1:
The patent extends single-dimensional event data mining to multi-dimensional analysis by incorporating machine identification attributes (such as machine type, model, serial number) as additional dimensions. This allows the system to detect patterns that span multiple machines while maintaining the efficiency of sequential pattern mining algorithms. The multi-dimensional approach enables differentiation between patterns specific to individual machines and those common across machine fleets.
2Reliability
If multi-dimensional sequential pattern mining is implemented, then common patterns across machine fleets can be detected, but the computational resources and processing time increase
Solution Approach 1:
The patent segments the data mining process into two stages: first, extracting sequential patterns from event data using efficient algorithms; second, filtering and analyzing these patterns through machine identification attributes. This segmentation allows the computationally intensive pattern extraction to be performed independently, then filtered by machine attributes to identify cross-machine patterns, reducing overall processing time while maintaining reliability.
3Adaptability or versatility
If event data from multiple machines is collected and analyzed, then comprehensive pattern recognition is achieved, but the data volume and storage requirements increase
Solution Approach 1:
The patent extracts only the necessary information for pattern recognition by incorporating machine identification attributes (machine type, model, serial number) alongside event data. This selective extraction allows the system to maintain comprehensive monitoring capability across machine fleets while reducing data volume by excluding unnecessary machine-specific operational details that do not contribute to cross-machine pattern recognition.
Data Source
AI summary
The present disclosure is related to a method for monitoring at least one event data generating machine, including a data logging device for providing event data. The method comprises transferring logged event data from at least one of the event data generating machines to a central processor, mining a multi-dimensional sequential pattern within said transferred event data wherein at least one dimensional attribute holds information indicating said event data generating machine or the at least one event data generating machine property, and matching said mined multi-dimensional sequential pattern with patterns stored in a central pattern database.