Graph-Based Alarm Handling for Faster Root Cause Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Operators in industrial plants face challenges in quickly obtaining a comprehensive overview of the status and topology of processing systems, especially during critical phases or alarm situations, due to scattered information across multiple screens and data sources, leading to time-consuming and error-prone alarm analysis.
Innovation Solution
A method that represents the processing system as a directed graph, allowing selection of a starting node to construct a subgraph, outputting relevant nodes and attributes, and using machine learning for anomaly detection and visualization to provide a clear, interactive overview of the plant's state, including alarm-related information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If operators manually analyze scattered alarm information across multiple screens and data sources, then comprehensive alarm analysis can be achieved, but time consumption and error probability increase significantly
Solution Approach 1:
The patent merges scattered alarm information from multiple screens and data sources into a unified alarm analysis interface. The system integrates process data, alarm history, and diagnostic information into a single comprehensive view, eliminating the need for operators to manually switch between multiple screens and data sources, thus reducing time consumption while maintaining analysis accuracy
Solution Approach 2:
The patent introduces an intelligent alarm analysis system as an intermediary between the complex processing system and operators. This intermediary automatically collects, processes, and presents relevant alarm information, reducing the manual effort required while improving analysis accuracy through systematic data integration and presentation
2Reliability
If operators manually investigate complex multi-device equipment failures, then root cause analysis can be performed, but time and costs increase due to the complexity of the investigation
Solution Approach 1:
The patent implements preliminary action by pre-configuring diagnostic rules, relationships between process elements, and analysis algorithms before failures occur. When a failure happens, the system immediately applies these pre-prepared diagnostic frameworks to rapidly identify potential root causes, eliminating the need for operators to manually investigate complex multi-device failures from scratch
Solution Approach 2:
The patent incorporates feedback mechanisms where the alarm analysis system continuously monitors system state, compares actual performance against expected behavior, and provides feedback about potential root causes. This automated feedback loop enables rapid identification of failures and their causes, improving both reliability and response speed
3Loss of information
If comprehensive process data and topology information are displayed to provide complete overview, then understanding of plant status is improved, but information complexity and difficulty of detection increase
Solution Approach 1:
The patent applies segmentation by dividing comprehensive process data and topology information into organized, hierarchical groups based on process areas, equipment types, or alarm severity. This segmentation allows the complete plant status to be displayed in a structured manner that reduces analytical difficulty while maintaining information completeness
Solution Approach 2:
The patent implements local quality by providing different levels of information detail and presentation formats for different process areas or alarm types. Critical alarms receive prominent display with detailed diagnostic information, while normal operations show summarized data, optimizing the balance between information completeness and analytical ease for different system states
Data Source
Figure 1
Figure 2~3
AI summary
The invention relates to the field of alarm handling or for providing an attribute of an element in an industrial plant or a processing system, particularly for processing systems that can be represented as a graph. For this, the processing system is represented as a directed graph comprising a plurality of nodes and directed edges, wherein each node represents an element, each node has an attribute, and each directed edge represents a relation between two elements of the plurality of elements. The method comprises the steps of: selecting one node of the plurality of nodes as a starting node; constructing a subgraph, wherein the subgraph consists of all the nodes that are forward-connected by at least one directed edge from the starting node; and outputting all nodes and the attribute of the nodes of the subgraph.