ABSTRACT
In recent years, analysis methods using sensor data and record data acquired through the provision of maintenance services for social infrastructure equipment have attracted considerable attention. We focus on optimal replacement of elevator components. Features of elevators include that they move continuously for a long time without any operator and their proportion is high among social infrastructure equipment. We define five steps in the analysis of maintenance services for social infrastructure equipment. We have developed an analysis system consisting of life-limited component analysis, replacement planning simulations, and service performance analysis. The analysis system uses a combination of functionality of machine learning, such as ontology processing, text mining, and facility-type clustering in order to handle various types of facility data.
