Mechanical equipment offline detection system vibration signal spectrum analysis and alarm method

Abstract

The invention discloses a mechanical equipment offline detection system vibration signal spectrum analysis and alarm method comprising the following steps: performing Fourier transform on a vibration signal to get a vibration amplitude sequence and a frequency component sequence; using a Gaussian function template to smooth the spectrum data of the vibration signal to get the amplitude envelope of the whole band spectral line, taking the smoothing length as an adjustment parameter, setting an alarm threshold, and taking the ultra-limit total value as an alarm parameter; performing adaptive frequency band division, automatically adjusting the obtained number of frequency bands based on the spectrum complexity of data, and setting a relative alarm threshold for the vibration energy value of each frequency band on the basis of the vibration energy value of each frequency band so as to identify whether the distribution of vibration energy of a newly generated vibration signal is abnormal and to automatically identify whether the state of mechanical equipment is abnormal. The method overcomes the defects of traditional vibration signal spectrum analysis, improves the ability of automatic vibration signal spectrum analysis and the accuracy of alarm, meets the security requirements of mechanical equipment operation and ensures safe and reliable operation of mechanical equipment.

Description

technical field [0001] The invention relates to a vibration signal frequency spectrum analysis and alarm method of an off-line detection system for mechanical equipment. Background technique [0002] In the off-line detection system of mechanical equipment, Fourier transform is performed on the signal detected by the vibration sensor, and the functional relationship between the vibration signal and the frequency is established with the frequency as the independent variable, which is called spectrum analysis or frequency domain analysis. Spectrum analysis decomposes complex time signals into several single harmonic components for research, and obtains the frequency structure of the signal as well as the amplitude and phase information of each harmonic, which is the most commonly used and most important analysis in mechanical equipment condition monitoring and fault diagnosis It is one of the methods, and it is also the core diagnostic tool in the off-line detection system of ...

AI Technical Summary

Problems solved by technology

[0005] At present, the existing spectrum automatic alarm method is to set the threshold of the frequency component amplitude or frequency band energy value corresponding to the typical fault. This alarm method is greatly affected by the working conditions. Under normal conditions, the frequency spectrum structure of the vibration signal also fluctuates to a certain extent, which makes it difficult to locate the monitoring frequency components and frequency bands, causing false alarms and missed alarms

At the same time, it is closely related to the spectrum structure, and it is more effective for displacement signals with little interference and few spectral lines. However, the most commonly used speed and acceleration signals in the production site have a poor alarm effect due to their complex frequency components.

Secondly, the alarm method is aimed at the local characteristics of the frequency spectrum, but mechanical equipment is a complex system composed of multiple parts, and the characteristic frequency monitoring is carried out for each fault of each part under each working condition, which cannot be realized in the application

[0006] Take Commtest's Ascent vibration monitoring and diagnosis software as an example. It mainly has two functions of spectrum envelope and frequency band alarm in terms of spectrum alarm, but its spectrum envelope only provides simple statistical envelope calculation and cannot reasonably describe the full frequency band. The distribution characteristics of the inner spectral line, and the frequency band alarm setting is completely dependent on manual adjustment

Images