Intralogistics Fault Source Identification Using Dynamic Graph Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern intralogistics systems are complex and challenging to diagnose faults in due to their high complexity and numerous interacting components, leading to inefficient fault determination methods that rely heavily on operator experience rather than technical models.
Innovation Solution
A computer-implemented method using a graph model to determine fault causes by analyzing signal sequences from various components, including sensors and machine learning algorithms to identify anomaly patterns and cause-effect relationships, reducing reliance on operator experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If a static model (construction plan or parts list) is used to map the intralogistics system, then the system structure is documented, but the model cannot reliably identify fault causes due to complex and undocumented functional interactions
Solution Approach 1:
The patent transforms the static construction plan into a dynamic graph model that continuously updates based on actual system operations and fault data. The graph model evolves by incorporating new cause-effect relationships discovered during system operation, allowing it to adapt to complex functional interactions that cannot be predetermined in a static documentation.
Solution Approach 2:
The system automatically discovers and documents causal relationships between components by analyzing operational data and fault patterns without requiring manual documentation of every functional interaction. The graph model self-updates with new relationships identified through machine learning algorithms, eliminating the need for exhaustive manual system documentation.
2Productivity
If fault analysis relies on operator experience rather than technical models, then practical fault identification is achieved, but the method is not scalable and effective for complex systems
Solution Approach 1:
The patent replaces the mechanical reliance on operator experience and manual fault analysis with an automated computational system. Machine learning algorithms analyze system data and the graph model to automatically identify fault causes, substituting human expertise with an automated technical solution that scales to system complexity.
Solution Approach 2:
The system implements continuous feedback loops where fault analysis results are fed back into the graph model to refine and improve future fault identification. The machine learning algorithms learn from each fault incident, progressively improving the accuracy and efficiency of automated fault determination through iterative optimization.
3Reliability
If the intralogistics system operates with high availability requirements, then system performance is maintained, but fault conditions cause lengthy downtimes due to slow cause determination
Solution Approach 1:
The system performs preliminary actions by continuously maintaining an updated graph model of cause-effect relationships during normal operation. When a fault occurs, the pre-computed causal pathways enable rapid identification of root causes without requiring time-consuming analysis during the fault condition, thus minimizing downtime while maintaining high availability.
Data Source
AI summary
A method determines a cause of fault in an intralogistics system, wherein functional cause-effect relations between components are recognized from signal sequences of components of the system during anomaly situations and mapped in a graph model. Based on this graph model, one or multiple probable causes of a current anomaly of the intralogistics system are computed. It can be provided that, for creating the graph model, already known fault patterns are additionally taken into account and/or a manual input of causes of fault for an anomaly is done via an operator terminal.