Modern optimal control theory is applied to the control of the ram velocity of molten polymer in the mold filling process of a injection molding machine. Dynamic equations of the filling process are formulated, taking into account the nonlinearity of hydraulic control system and the non-Newtonian flow of molten polymer in the cavity. Several unmeasurable coefficients included in the equations are determined by the parameter optimization, based upon the minimization of errors between the measured and predicted response, using the Steepest descent method.
The nonlinear optimal control problem of mold filling process leads to an open loop controller. But to make the most of the merits of it, closed-loop control scheme is adopted. Using the Direct search method, the optimal parameters for the constant gain feedback controller are determined.
The control objective is to minimize the errors of ram velocity and desired ram velocity profile vs. ram stroke in the mold filling process without the excessive control energy expenditure, where the desired ram velocity profile is selected for reduction of the residual stress of the products to be expected. In reality, all of the state variables cannot be measured. So, output feedback controller is implemented to satisfy present control objective.
For experiments, a microcomputer is interfaced with a conventional injection molding machine, servovalve and appropriate sensors. And machine level language programs are developed to be applied to the control of mold filling process and data sampling.
