Multi-agent Communication and Collaboration of Aero Engine Health Management

Abstract

For the four processes of aero engine health management, which are composed of condition monitoring, diagnostics, prognostics and prescriptive action, a four-layer architecture of aero engine health management divided into data layer, transaction layer, decision layer and users layer, is designed based on multi-agent. Corresponding to the function of each layer, six agent roles are defined, such as condition monitoring, information fusion, fault diagnosis, task consultation, rules management and human-computer interaction. FIPA ACL(Agent Communication Language)is the foundation for multi-agent interaction. The various steps involved in developing a multi-agent system using JADE(Java Agent Development Framework)are discussed. This paper provides theoretical conception for intelligent monitoring and intelligent fault diagnosis of aero engine, and for further research of multi-agent collaboration on multi-agent system.