| Co-Investigator(Kenkyū-buntansha) |
HASHIZUME Hidetoshi Graduate School of Engineering, Tohoku University, Professor, 大学院・工学研究科, 教授 (80198663)
TAKAHASHI Makoto Graduate School of Engineering, Tohoku University, Research Associate, 大学院・工学研究科, 助手 (00243098)
|
|---|
| Research Abstract |
The present research project was launched in 1997 aiming at improving the reliability and safety of large-scale artifacts such as a nuclear power plant, space shuttle, aircraft, etc. The main idea behind the project is the belief that the improvement of preventive maintenance, with particular emphasis on the condition-based maintenance with a broad perspective covering the whole plant lifetime, is the central issue to meet the need.
Furthermore, base on the fundamental recognition that the essential nature of the required database technique for condition-based decision-making is the intelligent interface, efforts were centered in development of the three key-techniques described below :
(a) Lifecycle database with evolutionary adaptability for condition-based maintenance support,
(b) Mobile-agent implementation of monitoring and diagnosis tools,
(c) Goal-oriented sensing for intelligent information acquisition.
In the subproject-(a), an efficient data search scheme was introduced by utilizi
… More
ng a specifically introduced similarity indices called partial and heirarchically-defined similarity index. The search scheme was found to be used as a technical basis decision-support system for maintenance problem solving. The high modifiability of the database was also confirmed to be useful for evolutionary development of the database. In the subproject-(b), the monitoring and diagnosis tools were developed and implemented by the mobile-agent technique, which allowed dynamic task allocation and evolution to accommodate computational load focusing and technology innovation.
The essence of technical development in the subproject-(c) is the goal-oriented sensing, which implied cooperative use of conventional and mobile sensing systems. The locations of the mobile sensor were determined based of the measurement by the conventional sensing system and the causal reasoning based on the model called Bayesian net. The high potentiality of the proposed techniques as essential constituents of the condition-based maintenance support system for ensuring life cycle integrity was clearly demonstrated through the study. Less
|
