Operational Safety Assessment of Turbo Generators with Wavelet Rényi Entropy from Sensor-Dependent Vibration Signals

Abstract

With the rapid development of sensor technology, various professional sensors are installed on modern machinery to monitor operational processes and assure operational safety, which play an important role in industry and society. In this work a new operational safety assessment approach with wavelet Rényi entropy utilizing sensor-dependent vibration signals is proposed. On the basis of a professional sensor and the corresponding system, sensor-dependent vibration signals are acquired and analyzed by a second generation wavelet package, which reflects time-varying operational characteristic of individual machinery. Derived from the sensor-dependent signals’ wavelet energy distribution over the observed signal frequency range, wavelet Rényi entropy is defined to compute the operational uncertainty of a turbo generator, which is then associated with its operational safety degree. The proposed method is applied in a 50 MW turbo generator, whereupon it is proved to be reasonable and effective for operation and maintenance.

Fault Diagnosis of Demountable Disk-Drum Aero-Engine Rotor Using Customized Multiwavelet Method

“…Moreover, it must be said that although a similar multiwavelet method is also be used for equipment condition identification in reference [ 13 ], the severe shortcoming of low adaptivity on the developed multiwavelet basis greatly limits the ability of customized lifting multiwavelet on signal feature extraction. In addition, some findings need to be introduced based on the experimental results.…”

“…As a powerful tool for describing the non-stationary signal, wavelet transform (WT) [ 9 , 10 ] has already shown its tremendous effectiveness in mechanical equipment condition monitoring and fault diagnosis because of its property of multi-resolution analysis [ 11 , 12 , 13 ]. Different from Fourier transform, a specific fault symptom can be matched and extracted by WT on the basis of selecting the basis function from the basis function library, which is greatly beneficial to fault feature extraction.…”

“…The performance of any existing wavelet basis function from the basis function library can be enhanced by the lifting scheme according to the actual requirement in the special engineering application, which brings about the possibility of constructing a customized wavelet. Chen et al [ 13 ] developed customized lifting multiwavelet based on the common lifting scheme for the condition identification of mechanical equipment. However, the mentioned method has the severe shortcoming of low adaptivity on the developed multiwavelet basis, which limits the ability to customize lifting multiwavelet on signal feature extraction.…”

“…Lin et al [ 22 ] identified the misalignment fault of the motor shaft based on multi-scale entropy for wavelet denoising. Many kinds of definitions for information entropy have been developed in detail for some special and concentrated applications, such as fuzzy entropy [ 23 ] for customized bacterial foraging, sample entropy [ 24 ] for the tool condition detection, hierarchical entropy [ 25 ] for biological data analysis, cross entropy [ 26 ] for the multi-target tracking, Rényi entropy [ 13 , 27 ] for high-resolution scalar quantization, and so on. Since wavelet transform has the powerful ability to describing un-stationary vibration data in both the time and frequency domains.…”

Condition Monitoring for Roller Bearings of Wind Turbines Based on Health Evaluation under Variable Operating States

“…The operational safety degree was increased in Reference [48] and the Renyi entropy defined as: signals to the NI-9401, where one represented the torque signal in an impulse sequence for a specific frequency, and the other the rotation pulse from the optical encoder in the tachometer. Details on the data acquisition system are shown in the Appendix A.…”