Development of a system for corrosion diagnostics and prognostics

Typically, the majority of the cost of maintaining aging aircraft structure can be associated with corrosion prevention and control. The major portion of the cost of inspecting aircraft for corrosion damage is associated with obtaining access to hidden parts of the airframe. In addition, there is the added cost associated with incidental damage that is done to the structure while gaining access to the hidden areas. There is a clear need to develop in-situ sensors and the diagnostic and prognostic algorithms to monitor corrosion environments to provide early warning of the onset of corrosion within hidden areas of aircraft. This paper covers our proposed approach to the development of an intelligent health management system. The system is developed around the principal of network intelligence based on autonomous sensing agents (provided by a distributed agent-based network and an intelligent object-based IT platform) closely linked to multi-functional, multi-stage sensing. In principle, prognostics will be embedded in different parts of the network. This will lead to significant advantages, not least of which is a higher level of accuracy provided by network learning and the ability to respond directly to sensor input. The system involves a multi-sensing approach and requires correlating sensor readings with actual measured corrosion (maximum pit depth). The general concept of the system will be presented, and the strong correlations found between sensed damage indicators and real surface damage for two aluminium grades (7075 and 2024) will be discussed.