Model-based approach to real-time monitoring of machine health condition

The semiconductor industry is one of the most technology-evolving and capital-intensive market sectors. To strengthen a competitive position in this industry, the key challenges are the increase of production throughput and yield. Effective prediction of the failures of the semicon equipment is a way to decrease a machine downtime, improve productivity, and reduce production costs and repairing time.

ASM Pacific Technology developed an infrastructure for condition monitoring of a generic motion stage of a semicon machine. This infrastructure makes sure the digital twin of our machines reacts as realistic as possible. It is based on the first principle models of the stage power electronics, three-phase motors, mechanics, and control. The original nonlinear continuous-time models are discretized and implemented in a state-observer for real-time estimation of different machine signals, especially the ones that are not directly measurable. By monitoring these signals, we can detect deviations of the machine performance from the nominal one. That helps us determine root-causes of these deviations and optimally schedule the machine maintenance.


About the speaker

Robin van Es is a control engineer at the ASM Pacific Technologies (ASMPT) Center of Competency in Beuningen. He holds a MSc degree in mechanical engineering received at the Eindhoven University of Technology.
He works on the problems of multi-disciplinary machine modeling, simulation, control technology, mechatronics system development and optimization, as well as machine health management.