基于相空间重构技术的断路器振动信号边缘特征提取

Mechanical fault is one of the main faults of circuit breaker，and the vibration signal generated during the opening and closing period of circuit breaker can effectively reflect its operating condition. Aiming at the limitation of the existing linear analysis methods such as time-frequency domain analysis in extracting circuit breaker vibration signals，a kind of method for extracting edge characteristics of circuit breaker vibration signals based on phase space reconstruction is proposed and discussed，which is used to characterize the nonlinear characteristics of circuit breaker and provide a basis for the condition diagnosis of circuit breakers. The fault simulation platform is set up to simulate the vibration signal of the normal condition，contact wear，coil aging，low operating voltage and closing spring failure. The edge image and edge characteristics of the vibration signal is extracted in combination with fractal theory and edge detection algorithm，and the effectiveness of the edge characteristics on the vibration signal diagnosis is analyzed from such three perspectives as the shape of the edge image，the value of the edge feature and the energy of the reconstructed circuit breaker power system. The influence of different edge detection algorithms and spatial dimensions on edge feature extraction is studied. The study results show that ，under specific spatial dimension，the extracted edge characteristics can distinguish mechanical conditions of different circuit breakers and have certain application prospects.