Dynamic Alarm Suppression via Failure-Cause-Effect Graph
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional alarm systems fail to dynamically suppress redundant alarms concurrently with their generation, leading to a large number of alarms being displayed, making it difficult for users to identify and react to root failures, especially in systems experiencing multiple concurrent failures.
Innovation Solution
An alarm system comprising a process control unit, a fault diagnosis and isolation (FDI) unit, and a filter that generates a failure-cause-effect graph to dynamically identify and suppress redundant alarms, while displaying non-redundant and not-yet-classified alarms, using this graph to classify alarms as causes or effects of previously diagnosed failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If conventional alarm systems display all generated alarms, then complete information is provided to the user, but the large number of redundant alarms makes it difficult to locate important alarms and react to root failures
Solution Approach 1:
The alarm system segments alarms into different categories (redundant vs. important) using a failure-cause-effect graph. The graph structures alarm relationships by identifying root causes and their effects, allowing the system to separate and prioritize alarms for display based on their causal significance rather than displaying all alarms equally.
Solution Approach 2:
The failure-cause-effect graph acts as an intermediary between alarm generation and alarm display. This graph structures the causal relationships between alarms and enables intelligent filtering, allowing the system to suppress redundant alarms while preserving important ones, thus mediating between complete information availability and ease of operation.
2Ease of operation
If hard coded alarm suppression rules are used to suppress redundant alarms, then the number of displayed alarms is reduced, but multiple concurrent failures may be hidden and the system requires cumbersome setup and maintenance
Solution Approach 1:
The system transitions from static hard-coded suppression rules to a dynamic failure-cause-effect graph that is continuously updated based on diagnosed failures. This dynamic structure adapts to different failure scenarios and concurrent failures automatically, eliminating the need for manual rule configuration while maintaining reliability in detecting multiple simultaneous failures.
Solution Approach 2:
The failure-cause-effect graph automatically updates itself based on failure diagnoses without requiring manual intervention for rule setup or maintenance. The system self-configures the alarm suppression logic by dynamically constructing causal relationships from diagnosed failures, eliminating the cumbersome setup and maintenance of hard-coded rules.
3Device complexity
If time stamps are used to suppress alarms occurring at later times, then redundant alarms are suppressed, but the system requires high time resolution and highly synchronised system clocks
Solution Approach 1:
The system replaces the mechanical time-based suppression mechanism with a causal relationship-based mechanism using the failure-cause-effect graph. Instead of relying on precise time stamps and synchronized clocks to determine alarm suppression, the system uses logical causal relationships derived from failure diagnoses, eliminating the need for high time resolution and synchronization.
Data Source
AI summary
The present invention is concerned with an alarm system and a method for suppressing redundant alarms in a monitored system. The alarm system comprises a process control unit, an FDI unit, a filter, and an alarm display unit. The process control unit detects anomalies in the monitored system and generates alarms corresponding to them. The FDI unit diagnoses failures in the monitored system and generates a failure-cause-effect graph comprising a list of generated alarms, along with the alarms corresponding to the causes and effects of each listed alarm. The FDI unit dynamically updates the failure-cause-effect graph concurrently with the generation of alarms by the process control unit and the diagnosis of failures by the FDI unit. The filter receives the generated alarms and identifies redundant and not-yet-classified alarms from the received alarms, by using the failure-cause-effect graph. The filter suppresses the identified redundant alarms and passes on non-redundant and the not-yet-classified alarms to the alarm display unit, for display.