Optimization of a vacuum thermal evaporation process through Model-based Predictive Control of the source temperature. Issue 11 (2015)

• Record Type: Journal Article
• Authors: Zöller, D.; Reiter, M.; Abel, D.
• Is Part Of: IFAC-PapersOnLine. Volume 48:Issue 11(2015)
• Page Start: 86
• Page End: 91
• Publication Date: 2015
• Subjects: Model-based control -- physical model -- predictive control -- process parameter estimation -- production systems -- temperature control; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2015.09.164
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: The Vacuum Thermal Evaporation (VTE) process is a technique for the production and deposition of thin films, which are used in various industrial applications. Here, the coating process is performed by evaporating a raw material in a high vacuum (HV) environment. The major goal is to produce a thin layer through a well-defined deposition rate. Due to high demands on the layer thickness, the tolerance for the temperature within the vacuum chamber is less than 0.2 K. This requieres the design of very well tuned controllers. In this paper, a Model-based Predictive Control (MPC) approach is presented for controlling the temperatures in the VTE process that explicitly takes into account the limitations arising from the requirements. First, the heat transfers within the heater and the source are investigated and described by physically motivated equations. As the heat transfer in the HV is dominated by radiation, this approach leads to a nonlinear system. Reliable measurement data are used to parameterize the previously derived dynamic model of the temperature behavior. An advantage of this physically motivated model is, in contrast to a linear operating point model, to allow an extrapolation and therefore to adequately depict the system behavior in a large operating range. Based on this approach, a MPC is developed and implemented on a pilot plant. It is shown that the high requirement on the temperature accuracy is met by the use of the proposed MPC. Experimental results show the potential of the introduced control scheme.