Mixed-Effect Model of the Fluid Viscosity for Virtual Sensing of the Flow-Front Dynamics

For online control and fault monitoring of the epoxy infusion in a vacuum assisted resin transfer moulding (VARTM) process, good knowledge of the current state of the flow-front is essential. Due to heterogeneous material properties and the environmental conditions, the permeability of the medium, as well as the viscosity of the epoxy, can change significantly during the epoxy infusion process. Hence, for a fast and reasonably accurate estimation of the flow-front, a virtual sensing system capable of combining the physics of the system and the measured data is needed. In this short paper, we propose a data-driven mixed-effect model for the fluid viscosity data acquired from multiple experimental. The proposed model can be easily integrated into the stochastic differential equations (SDE) based virtual sensing system for flow-front dynamics in a VARTM process.