Prognostics and health management solution development in LabVIEW: Watchdog agent® toolkit and case study

Abstract

Prognostics and Health Management (PHM) research has been intensively studied in various industries in order to improve the overall performance of the critical assets and avoid unexpected failure. For data-driven approach, many data analysis methods have been applied for PHM system development including digital signal processing, data mining, machine learning and pattern recognition. And three major goals for PHM system, including health assessment, fault diagnosis and remaining useful life prediction, are achieved by combining different algorithms. However, the challenges for industry are how to efficiently select appropriate tools to develop a suitable PHM solution and how to quickly demonstrate PHM concepts for different applications. The concept of a reconfigurable algorithm toolset titled the Watchdog Agent® for PHM was first presented in 2003 and now is a commercialized toolbox in LabVIEW. The Watchdog Agent® toolbox consists of selected tools/algorithms from four categories: signal processing, health assessment, fault diagnosis and remaining useful life prediction. LabVIEW, a system design software developed by National Instruments, has been used for measurement, testing, control and data analytics in various areas including wind energy, automobile manufacturing, aerospace, etc. This paper first presents an introduction about Watchdog Agent® (WDA) Toolkit and then provides a systematic approach for PHM solutions development in LabVIEW environment. A detailed discussion about data acquisition, data pre-processing, feature extraction, model training and result visualization are provided with case studies.