Research Of Modeling And Learning Control Of Die Cavity Pressure In Hydrodynamic Deep Drawing Equipment

Hydrodynamic deep drawing (HDD) is a very practical sheet forming technology and one of the advanced technologies for manufacture and it has a bright future for industry application, because of its guaranty on quality of reduction of cost and erection time. In the process of HDD, cavity pressure control accuracy is a key factor for improving sheet forming limit.The system contains many nonlinearities, such as vavle deadzone, valve flow saturation, valve's nonlinear flow characteristics, the stick-slip friction in the cylinder and a significant compliance due to the hose connector between the hydraulic parts. These affect the control performance with the dead zone at the start, bigger overshoot and longer settling time in the dynamics and the obvious output error in steady state, all of the factors weaken the advantage of the hydrodynamic deep drawing. How to choose the proper control strategies to improve the system control performance so as to getting good product quality is the original intention of the research.Firstly, this paper analyzes the specialty of HDD system in detail, then aiming at the high nonlinearities, establishes the nonlinear model of the cavity pressure control system by power bond graph, which can be applicable in large-scale, and examines the validity of the model by comparing with the experimental data.Because the deep drawing process is the same at each cycle and the given trajectory of the cavity pressure is identical at the same patch of product, Iterative Learning Control (ILC) is selected as the control strategy of this system. The ILC works well when the information of structure and parameters of the process is unavailable. The simulation results show the merit of ILC in reducing dead zone and restraining overshoot and output error in steady state. Finally, the thesis examines the feasibility and validity of the then examines the feasibility and validity of the arithmetic.