Study On The Process Controland Strip Performance Of Vibratory Twin-Roll Thin Strip Casting

The technology of twin-roll strip casting is one of the shortest hot-rolled steel strips technologies in the world nowadays.It is a green production process,which has the natural advantages of resource saving,low energy consumption,low emission,environmental friendliness,low investment and low operation cost.Although the twin-roll strip casting technology has been widely valued in the world steel industry,it still needs to be improved to realize industrialization as soon as possible.In the casting process,the solidification time of liquid steel from entering molten pool to casting and rolling through crystallization roller is very short,so the controllable range of process parameters is very narrow.The minimum fluctuation of parameters may adversely affect the strip quality.Therefore,it is required to monitor accurate process parameters,and adopt the proper strategy to precisely control the process parameters,so as to ensure the stability of the casting and the quality of the thin strip.In this paper,the influence relationship between technological parameters in the process of the vibratory twin-roll thin strip casting,the development of rolling force model,process monitoring system,the control strategy of technological parameters and the design of control system are studied.The results have important theoretical significance and practical value for the process control of vibratory casting.Based on the theory of the heat transfer,fluid dynamics,solidification and dynamic grid technology,aiming at the characteristics of vibratory casting process,the fluid dynamics finite element analysis software Fluent is used to conduct the heat flow coupling numerical simulation of metal liquid in the molten bath.Then,the influence law of casting speed,casting temperature,liquid level height,roll gap size and vibration condition on Kiss point is obtained,which provides a theoretical basis for the research on control strategy of casting process.According to the heat transfer characteristics between rolling bath of the vibratory twin-roll thin strip casting and the contact of the crystallization roll,the contact of the vibrating bath is divided into the casting zone and the plastic deformation zone with theKiss point as the boundary.The deformation mechanism of metal in plastic deformation zone is analyzed by vibratory casting-rolling according to the front slip zone,the back slip zone and the plastic deformation zone.Based on the analysis results and finite element analysis theory,the finite element stress analysis of metal in rigid contact zone is carried out by trapezoidal micro-element,and the rolling force model of vibratory casting is established by using hot rolling theory.Finally,the model of casting-rolling force is validated with experimental data.The casting and rolling process of vibratory twin-roll thin strip casting and the characteristics of the equipment itself are analyzed.On this basis,the hardware scheme of the monitoring system for casting and rolling process parameters is worked out.The control program of upper computer and lower computer is realized by using Labview and Step7 software platform.The monitoring system of ?500×350 vibratory twin-roll thin strip casting is established,which provides accurate?fast and reliable data support for the research of casting process,and provides a platform for the realization of control strategy.The control strategies of the process parameters of vibratory casting are analyzed,and the control strategies of HAGC,AFC and Kiss point are formulated.The corresponding control strategies are put forward for the casting start stage and the stable stage of casting.Radial Basis Function(RBF)Neural Network is selected as the research method.In order to improve the training speed and generalization ability of RBF Neural Network,the training algorithm of network parameters is improved.It is determined that the casting control system adopts NARX model class for non-linear modeling,and the training model order of neural network is determined by AIC criterion.The accurate nonlinear system identification of Casting HAGC system and rolling system force is carried out by using Simulink,and the identification accuracy was verified by experimental data.Based on the neural network model of the control system identified by the system,the Indirect neural network self-correcting control structure is designed,and the self-correcting control algorithm is improved.The design of HGAC system and casting force control system of the casting-rolling mill is completed.The two intelligent control systems and the conventional PID control system are simulated by Simulink.The comparison shows the superiority of RBF neural network in solving the problems ofnonlinearity and uncertainty in the servo control system.The indirect RBF neural network self-correcting control modules of two control systems are realized by using Labview software platform,which is integrated in the ?500x350 vibratory twin-roll thin strip casting monitoring and control system.The conventional rolling experiment and casting experiment on vibratory casting mill verify the intelligent control system.On the experimental platform of ?500×350 vibratory twin-roll thin strip casting,a comparative experimental study on non-vibratory casting and vibratory casting of 20 CrMn steel at different casting temperatures was carried out.The experimental results show that vibration can effectively refine the grain size of the strip during casting,improve the liquid steel fluidity in the molten pool,restrain the shrinkage and promote the dispersion of fine second phase particles.Therefore,at different casting temperatures,the yield strength?tensile strength and elongation of the strips obtained by vibratory casting are much higher than strips obtained by the traditional non-vibratory casting.