Research Of Special Steel Bar Rolling Based On Automatic Gap Control

In the wake of intense competition in the market,people put forward higher request of bardimensional accuracy,mechanical property,production cost and development cycle.But bar rolling isa complex high-dimensional nonlinear process which demands that the coordination capacity ofeach rolling mill is extremely high.Front and back rolling mill status would be affected if fluctuationis produced in any mill.Although for the traditional bar rolling mill DC motor and looper device areused to achieve rolling without tension and torque,this can not guarantee the correct rollingdeformation and sectional shape everytime.According to the above problem,and insight into bar rolling technology,the Automatic gaugecontrol project presented in this article would be applied to bar continuous-rolling.As is known toall,AGC technology can improve very well dimensional accuracy of rolled piece,so it is widelyapplied in plate and strip rolling.According to the difficulties with AGC technology,this paper hasfocused on the following aspects:First,since high-dimensional deformation of bar rolling make force parameter detection becomedifficult,and rolling force is critical for AGC.Only by predicting rolling force which combines withspring equation and roll-pass system to set roll gap accurately,to improve bar dimensionalprecision。So according to the theories of engineering rolling force,a model of bar rolling forcebased on neural networks is founded.Next,a stationary model of position servo control system of hydraulic AGC is founded bystudying the key parts of hydraulic AGC system—servo-valve,hydraulic cylinder-flowequation,carrying capacity-balance equation and sensor,etc.Last,because of the inner loop of hydraulic AGC system—hydraulic automatic position controlis variable, strong coupling, non-linear,a hydraulic position controller would be founded based onmodel free adaptive control in the article. Through the deep learning model of free adaptive controltechnology,hydraulic position control program would be founded based on model of free adaptive control which exploit the control law to combine the estimation algorithm of pseudo-partial-derivative.At the same time MFAC controller will be founded by using MATLAB-SIMULINK.Bycontrasting PID and neural networks simulation, it turned out that MFAC controller has fastconvergence,small steady-state error,strong interference resistance and that hydraulic positioncontroller based on MFAC is effective.