Online Measuring Methods Of Mould Thermo-mechanical Behaviors And Their Application In The Continuous Casting

The mould is the core instrument for primary cooling and slab forming of liquid steel, and it takes effect as heat transfer, support, restriction, lubrication and demould during continuous casting (CC). The interior heat transfer and friction of mould have close relationship with the surface defects and breakout, which is very important for the effectively continuous casting. Aiming to control and adjust the initial solidification and frictional behavior online, it is necessary to ensure the conditions of stable heat transfer and lubrication, which is not only the requirement of strand quality, but also the precondition for stable and high efficient production. In the future, the development of continuous casting technology will more strongly depend on the automatic control during producing process. Therefore, it is urgently needed to develop new technologies for online measurement of thermal and mechanical behaviors in continuous casting mould, which is very important in both fundamental study and practical application. Furthermore, investigating on mould measurement methods and their applications, will play a significant role in further understanding of basic theory and monitoring practice for mould process.In this dissertation, fundamental studies of mould thermo-mechanical monitoring technologies and their applications are discussed, especially, the online measurement method of mould friction, as well as the numerical simulation of heat transfer and their applications. Based on the slab caster equipped with hydraulic oscillators, the theoretical model, measurement method and monitoring system of mould friction are investigated. In addition, according to the mould friction data measured on slab caster, the application method of mould friction is studied, and the prediction model and software for mould abnormalities are developed. In the last chapter of this dissertation, the calculation method, which combines the online measurement data of mould temperature and heat flux and numerical simulation, is investigated by using the inverse heat transfer model and neural network for the round billet continuous casting.Through theorecally analyzing the loading states during mould oscillation driven by the hydraulic units, the loading model is built and the theory foundation and overall measurement method are established for mould friction monitor in CC. The real-time calculation and optimization model of empty oscillating parameters is built by particle swarm optimization (PSO). The implementation and convergence performance of PSO algorithm are discussed, and the optimized results are also analyzed and evaluated. Based on the above work, both hardware and software for mould friction measurement are designed. The measurement tests show that the system is suitable for monitoring mould friction, and realizes the online detection of transient mould friction continuously, stably, and automatically. In the basis of the mould friction data measured, the periodical variations and characteristic values of mould friction, as well as negative strip parameters are discussed, and the characteristics of non-sinusoidal waveforms are also studied. In addition, by using the measurement data of displacement and cylinder force, the oscillation service state of hydraulic oscillators and its evaluation methods are investigated.According to the abnormal records of steel plant, mould friction measured on slab caster by power-method is used to investigate its response to abnormalities in continuous casting. The results show that mould friction can respond to most abnormalities such as breakout, submerged entry nozzle broken, and acute fluctuation of mould level. The root mean square of mould friction ranges from 0.8 to 4 kN before or during the abnormalities. The prediction model for mould friction abnormalities has been built by using artificial neural network models in combination with two pattern recognition algorithms. A set of software to predict the mould friction abnormalities in CC has been developed. The results of simulation prediction for online measurement mould friction data are found to be basically consistent with those collected from the abnormal records of steel plant, such as breakout, submerged entry nozzle broken and acute mould level fluctuations. In some cases, it can make prediction several minutes earlier than the alarm given by temperature breakout detection system. The proposed method can make the system respond fast enough in real time plant prediction, and render sufficient time for online prediction and taking operations against abnormalities.Based on the measured data of temperature and heat flux during round billet continuous casting, the calculation method which combines the online measurement data and numerical simulation is investigated. The thermal behavior is analyzed by an inverse heat transfer model, and it reflects the real state of transient heat transfer of mould. In order to decrease the calculation time to meet the online monitor requirement, an artificial neural network model is developed to solve the inverse problem, and the IHTP/ANN/DHTP (inverse heat transfer problem, artificial neural network, and direct heat transfer problem) integrated method of thermal behavior analysis from experimental temperatures is developed. The results show that the neural network is faster for inverse model, and the calculation results by this method can correctly reflect the characteristics of non-uniform heat transfer. Also the calculation accuracy and speed meet basically the requirements of online calculation. Through calculation, it is found that although the operating conditions are stable, the shell thickness may be fluctuated. The shell thickness varies directly with the heat flux at fixed mould circumferential degrees. It is also found that the profile of shell thickness is non-uniform and similar to that of the heat flux along circumference direction.