|
Abstract:
|
To identify and assess damage at the earliest stage is critical in manufacturing and aerospace industries due to the potential economic and life-safety consequences. The vibration data contain information about condition of the operating components. The goal of this research is to establish a new statistical framework for machinery health monitoring based on vibration data. First, formal statistical models will be proposed to describe the relationship between the vibration signals and some interesting factors, such as the damage level and environment settings. Second, damage detection and diagnosis are implemented based on statistical estimation and hypothesis testing. Though vibration is functional, novel methodologies will be needed because traditional functional data analysis might not work due to new features of data. The proposed methods will improve the power of damage detection.
|
