Data-Driven Condition Monitoring of AC-supplied Trip Coils of Circuit Breakers

Abstract

Efforts to achieve an uninterrupted and increasingly necessary economic power supply have led to condition monitoring of major components, including Circuit Breakers. This work focuses on monitoring a critical part of the circuit breaker: the solenoid, which triggers the opening and closing of the breaker’s contacts. A fault in this component often results in a malfunctioning breaker operation. The DC or AC supplied solenoid is part of the mechanical system that moves the contacts. Whilst monitoring concepts for DC-powered solenoids already exist and are extensively used in condition monitoring, this work introduces an approach for monitoring AC-powered solenoids by analysing their current profile. Due to the short activation time, transient processes must be considered. The AC current profile depends on the voltage phase at activation, which we demonstrate can be determined, without measuring the voltage. Using this information with training data enables prediction of healthy solenoid current curves. Profile analysis identifies mechanical events (e.g., anchor hitting end position, power cut-off) that are prominent in DC measurements but disguised in AC sinusoidal waveforms. Extracting these features from actual measurements and comparing it to the predicted curves identifies deviations and faults. An application example demonstrates that these features can reveal whether the operating voltage of the solenoid was within the permissible tolerance.