Fault Diagnosis of Elevator Based AR Bi-Spectrum

Abstract

According to the characteristics of the elevator fault vibration signals, proposing a Based on analysis of time series AR bi-spectrum elevator fault diagnosis. When the zero-mean, non-Gaussian white noise elevator device, the vibration signal using sampling to establish time series autoregressive model (AR model), resulting in AR bi-spectrum. Bi-spectrum signal processing is a new, powerful signal processing technology, which can be described the nonlinear coupling, suppression Gaussian noise and retention of phase information, you can get the elevator working status of the different dynamic characteristics. The results show that bi-spectral analysis with AR elevator failure is feasible and effective.

Virtual sensing-based monitoring and maintenance in elevators

“…Therefore, early detection of the aforementioned failures by means of condition monitoring have received major attention in the literature [23][24][25][26][27]30,32 in order to reduce the number of breakdowns.…”

“…Condition monitoring techniques employed in [23][24][25][26]32 are based on processing acceleration signals by means of classifiers such as artificial neural network (ANN), time-frequency domain techniques, wavelets and higher order spectra (HOS). These techniques are highly dependent on the quality of data, and furthermore, they employ accelerometers which are not readily available in the elevator.…”