무선 센서를 이용한 기계 건강 진단 및 예지를 위한 통합 소프트웨어

Abstract

This study presents integrated software for machine health diagnostics and prognostics based on measured signals using wireless unit sensors which have been developed in previous work. This software consists of four modules sequentially: data processing, data analysis and condition monitoring, diagnostics, and prognostics. Each module has its detailed function to identify faults from hidden signals. The algorithms which are used these diagnostics and prognostics modules have been developed and estimated from our studies in the recent time. In diagnostic algorithm, 5 classifier machines are combined for accurate decision of machine diagnosis. Additionally, some methods utilized to feature representation, feature selection, feature extraction, etc are also introduced in this study. The aim of developed software is to offer useful information and propose one of solutions for practical problems. In order to verify, two different data are used in this software, which are K-Water system pump data and 7 kinds of faulty simulated data. The results indicate that the software could be used as a reliability tool to real application.