The Research On Slagging Noise Monitoring Technology

Converter steel making is the dominant method in steel making technology. The process of slagging plays a very important role in steel making. Auto steel making needs the accurate and real-time estimation of the slagging status by computer technologies, as the foundation of the control system to adjust steel making for a stable process in fluency against slag splashing and drying; the loss of the iron water can be avoided and dust elimination and gas retrieving can be realized at the same time. Thus the efficiency and quality can be improved on the whole.With the exception of super converters, it is generally adopted that skilled workers make judgments according to their own experiences on the process and the condition in the converter. The lagging technique has been the formidable obstacle in converter automation. Many methods have been used to solve the problem in which the noise analysis is relatively better but it lack practicability in the absence of accuracy, stableness, and simultaneousness. Thus an appropriate method which can presents slagging condition accurately and conveniently is in urgent need with great potential of economy profits and social benefits.In this essay, the principle and work condition of the original audio slag controlling technology are analyzed, and the rationality and existing problems are pointed out with emphasis on the analysis of the causes, furthermore, the new slag-collecting system is designed and the new slagging noise signal processing method is gotten.This research with the theory analysis points out that the slagging noise is produced by stochastic impact, filtration, resonance and excitation which is superimposed by multi-stochastic process. On the basis of Matlab, with signal processing and the method of power spectrum analysis, contrasting varied different process, the multi-eigenfrequency band and its feature curve are determined; and the integrative eigenfunction which can describe the status of slagging is generated too. Furthermore, a filter is designed to eliminate the stochastic unstableness, meanwhile, some designations are given here to keep instantaneous reaction of the calculation and the feature curve more accurate at the same time.To realize the automatic recognition of the slagging condition, automatically recognizing system based on artificial neural networks is studied. It advances applying mask-off-code technology and sliding window technology to extract the feature of slagging noise and recognize it with BP neural network. This BP network has been demonstrated with high recognition capacity by sample training and testing.The process of steel making can be directed and stabilized using the slagging feature curve and recognition of slagging condition against slag splashing and drying.We plan to continue the study to improve the slagging recognition rate and the coordinated research on monitoring and controlling system to give the crucial theoretical support and technology preparation for automatically steelmaking process.