Eliminating External Factors with Variables Standardization for Monitoring Applications

Abstract

This paper proposes a novel preprocessing method for normalizing the measured variables of a system, with respect to external conditions. Our approach transforms the measured quantities into corrected ones that capture the internal behavior of the system while eliminating the impact of external variables on this behavior. We demonstrate the effectiveness of our approach through an experiment focused on vibration health monitoring in aeronautics. This preprocessing technique enables the use of consistent data for analysis and prediction across different operating conditions and thus enhances the accuracy and reliability of system monitoring.