Time-Frequency Analysis Tool for Intelligent Condition Monitoring Diagnostics

Abstract

Real-time condition monitoring is vital to prevent sudden failure and breakdown of critical power plant equipment. It leads to substantial financial loss due to service disruption, equipment damage, repair, and restart. Despite existing condition monitoring technology and adequate safety guidelines, many accidents have led to the sudden failure of power plant equipment, including electrical switchgear, in recent years. It is crucial to detect minor defects at the earliest to prevent them from turning into significant machine/equipment failures, which can lead to an unwanted outage in production and increase maintenance costs. An efficient condition monitoring technique can provide warnings and predict the faults at early stages by obtaining information about the machine in primary data. Currently, significant industries are relying on time-domain or frequency-domain analysis alone. The problem here is that these two approaches fail to yield accurate results for fault/transient signals due to their nonstationary nature. To overcome these limitations and obtain better information such as time of occurrence of specific abnormal frequencies using advanced single processing techniques, the authors developed an advanced Time-Frequency Analysis (TFA) Graphical User Interface (GUI) tool in MATLAB. This paper presents an innovative method to study condition monitoring both for offline and online analysis. It has been tested for robustness with fault data under normal as well as noisy conditions. The success of the proposed technique helps to develop an intelligent condition monitoring and diagnostic tool for intelligent health monitoring.