TIME-DOMAIN SIGNAL ANALYSIS FOR MONITORING OF ELECTROMECHANICAL SYSTEMS

Electromechanical systems often develop incipient faults: they are in a large number of types and their precise signatures depend on numerous parameters, including power supply, driven load and control circuit. Practical fault detection and diagnosis systems must exhibit high level of detection accuracy and acceptably low false alarm rates. Very common approaches are vibration monitoring (for detection of mechanical failures), motor current monitoring (for electromechanical faults) and monitoring of the negative sequence of motor current and impedance (for incipient detection of insulation faults). The flux monitoring also seems to be promising for diagnosis of electromechanical systems. The effectiveness of these various methods is reached or increased by stating whether the monitored signals own a particular global attribute (steady-state or not, modulated or not, disturbed or not, etc.) to obviously apply them the most suitable algorithm for signal analysis.