Research Projects

Framework for Anticipative Event Management and Intelligent Self-recovery

The project focused on the development of the core technologies for building Anticipative Event Management (AEM) systems with self-recovery capability. Anticipative event management represents a practical and novel approach for proactive management of manufacturing events. Key to the concept is the focus on predicting events that are likely to happen so that appropriate pre-emptive action can be taken, where possible, to prevent the undesirable exceptions from happening or worsening. It also emphasises on the need for self-recovery at system level based on the event characteristics and information available, so that the duration of an undesirable event occurrence is shortened, its scope confined, and its adverse impact minimised.


Contact PersonLuo Ming()
Solution

The project team has developed a set of core technologies for Anticipative Event Management (AEM) and self-recovery:

  • A methodology and a conceptual framework by which proactive and intelligent event management systems can be built: The methodology comprises four essential stages:Identification, Characterisation, Knowledge Formation, Design and Implementation. The first two stages are concerned with deriving and analysis of the requirements of the Anticipative Event Management (AEM) systems, while the latter two with the design and development of the system. The framework consists of four major function components:Event Monitor and Filtration (EMF), Intelligent Event Manager (IEM), Intelligent Self-recovery Engine (ISRE) and a Knowledge Base.
  • The core technologies for the rapid development of the four function components include a classification system for the analysing of manufacturing event characteristics, modified reference architecture for the development of Anticipative Event Management (AEM) systems, artificial neural network and G-value statistical based techniques to establish event co-relation models for the Intelligent Event Manager, and cost-based Genetic Algorithm and Ant Colony System for intelligent self-recovery engine.
BenefitsThe proposed methodology and framework have applied to a control of material handling system (MHS) in an industry case, in which modified hierarchical reference architecture for monitoring, predication and diagnosis of large automated systems have been applied. A prototype of self-recovery engine based on the computational intelligence has been developed and tested for the control of the MHS as well. The implementation and testing results from the real application demonstrated that the proposed methodology and core technologies are able to help industry to develop Anticipative Event Management (AEM) systems systematically and effectively.

Patents / Awards / Achievements / DifferentiationThe proposed methodology and framework have applied to a control of material handling system (MHS) in an industry case, in which modified hierarchical reference architecture for monitoring, predication and diagnosis of large automated systems have been applied. A prototype of self-recovery engine based on the computational intelligence has been developed and tested for the control of the MHS as well. The implementation and testing results from the real application demonstrated that the proposed methodology and core technologies are able to help industry to develop Anticipative Event Management (AEM) systems systematically and effectively.

ApplicationsAnticipative Event Management (AEM) applications include predicting equipment/component failures, performance degradation detection for proactive management of manufacturing operations.

Problems Addressed

Anticipative event management and intelligent self-recovery under the manufacturing environment have to address challenges relating to the areas of unobtrusive event abstraction, event co-relation and reconciliation, unsolicited exception processing, on-demand query, knowledge enriched self-recovery at system level, and multi-channel event notification and escalation.