Research On The Abnormal Vibration And Fault Monitoring For Heavy Thin-Slab Hot Rolling Mill

The CSP rolling mills introduced by Pearl River Iron Ltd reaches the 90's world advanced level. The violent abnormal vibration often occurs in the F3 rolling mill when it is rolling thin plates below 2.0mm, leading to 'Vibratory Lines' on the surface of its work rollers, and also on the surface of thin rolled steel plates. The roar at low frequencies arises because of varying vibration intensity. The violent vibration of the rolling mill speeds up the wear of its rollers, and affects its productivity and the quality of its products. So it is urgent to research the reason of vibration and its rule and put forward some methods for solving the problem.This paper aims at the problem of abnormal vibration in the rolling mill. After investigating on the spot and getting some fundamental information about the rolling mill vibration, we have made a number of field tests on the F3 rolling mill by employing a 12-channel data acquisition system. We have tested the vibration of the mill stand, the torque moment, bending moment, and the vertical and horizontal vibration of the work rollers, the vertical and horizontal vibration of the backup rollers, the vibration of the gear distribution box and the reduction gear and the rolling force. At the same time, we have also obtained the rolling processing parameters.After the time domain analysis, spectrum analysis, correlation analysis and amplitude cepstrum analysis, we ascertain that the vibration source of the transmission system of the rolling mill is its rollers. The vibration frequency ranges from 45Hz to 51 Hz.The basic frequency of torsional vibration of the F3 rollers ranges from 19Hz to 21 Hz.The separation between vibration lines on the rolled plates has a close relation to the basic frequency of the roller horizontal vibration. The dominant frequency of the roller vertical vibration is twice as large as that of the roller horizontal vibration. From these analytical results we have made some suggestions for rejecting vibration.Signal denoising lays foundations for further signal processing. In this paper we study the adaptive wavelet signal denoising method for rolling mill abnormal vibration signals, construct the parameter equations of orthogonal binary wavelet base, set up fitness functions of noise removal effect, adopt advanced genetic algorithm and realize adaptive wavelet denoising of rolling mill abnormal vibration signals. The wavelet denoising method works well for the rolling mill and hydraulic pump vibration signals.Due to the obvious non-linear characteristic of the abnormal vibration of the F3 rolling mill, especially the violent impulsive signals appearing at the moment when the rolling mill bites steel, in these cases, the analysis method based on FFT exists the limitation that cannot provide time-frequency analysis. In order to extract all frequency components from the signals appearing at the moment when the rolling mill bites steel and observe the concrete vibration form of all frequency components, we present the wavelet packet time-frequency analysis for the first time. We separate all frequency components from the signals and explain the starting process of the rolling mill abnormal vibration by the wavelet packet frequency band separation and...