Frequency Domain Analysis And Application Of Monitoring Signals In Continuous Casting Process

Continuous casting plays a key role in the whole steel making process,and is the core of structural optimization,quality improvement,variety development and energy saving consumption in steel industry.However,there are various kinds of quality defects such as breakout,bulging and so on,emerge one after another at higher casting speed,which has a great effect on the quality and production of steel.Therefore,the measurement of the important process parameters,exploring the diagnosis and prediction method for quality defects and process abnormalities are always the top priority for online detection in continuous casting process.In the past,when analyzing the signal,researchers usually only pay attention to the time domain characteristic of the signal,that is,the abnormalities are predicted based on trends and characteristics of signals over time.Due to the many factors affecting the monitoring signals in the production process,which leads to the weakening of the timing signal characteristics,and it is difficult to extract the useful information from the signals only through the time domain analysis.In this paper,the Fourier transform method is used to establish the time / frequency domain transformation model for the periodic detecting signals of continuous casting process.The measured data in the production process are analyzed in frequency domain,and the frequency domain characteristics of the signal are captured and discussed.The influence of casting parameters on the measured periodic signal is investigated.Through comparing process detecting signals under normal and abnormal conditions,the online prediction method for casting abnormity based on frequency domain transformation are further studied.Firstly,based on the frequency domain processing of the oscillating data and dynamic status of the hydraulic oscillation system,the frequency spectrum of the signal is obtained and studied by analyzing the base frequency,harmonic frequency,DC component and energy intensity in the frequency domain.The results shows that the frequency domain analysis can be applied to evaluate the symmetry and synchronic of the hydraulic oscillation system,as well as the influence of the casting parameters on the dynamic characteristics of mold oscillation.At the same time,the transient mold friction is analyzed by frequency domain analysis method.According to the analysis,the results show that the different oscillation state can lead to the change of friction density,and the frequency domain signal is sensitive to the abnormal change of mold friction.This research proves that it has potential application in the forecast of sticking breakout.After that,the model of time/ frequency domain transformation of mold level signal is established,and the discrepancy of frequency domain processing results for mold level at normal and abnormal status is investigated,and the results of frequency domain processing of mold level fluctuation are analyzed.To forecast and locate the bulging,combining with the equipment information of roller position and casting parameters,a prediction model for slab bulging based on frequency domain transformation of periodic mold level is put forward,and the feasibility of the model are investigated and verified.The results of this paper provide reference for the online detection and application of the periodic signals in continuous casting process.