Abstract
We study the state of production processes in detail to control and improve productivity and product quality. In this article we consider the problems on slow-response characteristics for heat-seal process. The slow-response processes are characterized by damped oscillation.
We modeled the boundaries of oscillating state and stable state and proposed the estimation problems on the first stable point of the stable state. This estimation by the method of least square is recognized as the problem on a nonconvex minimization.
To attain a globally optimal solution, we proposed a method for setting a starting point based on the model characteristics. Vast simulation results illuminated that our method converged to the preferable values without any negative variances using even any algorithms.
We also applied this method to practical data in the heat-seal process.
