Packaging Industry met reduced models

Printer-friendly versionEnviar a un amigo

The techniques of reducing models based on geometric modeling at multiple scales (coupling of differential models that work in different spatial dimensions) or modeling of reduced order (for example, adequate orthogonal decomposition or reduced base) have proved to be very solvent for the simulation of complex integrated systems, such as the human cardiovascular system.

The long experience of the Modeling and Scientific Computing Laboratory (MOX) in this type of models has been developed during the last years in a new industrial context, in the framework of an industrial collaboration with MOXOFF s.p.a. (Spinoff of Politecnico di Milano) and a world leader in the packaging of liquid foods.

The packages are assembled by the collapse of a papertube that, at the same time, is filled with the liquid product in a very fast and complex process. The design and control of this system, which must be carried out in an aseptic environment at high speed (up to 50000 packages per hour), is extremely difficult. In order to simulate the propagation of pressure waves through the cardboard tube filled with liquid produced by the collapse of the tube, a set of geometric models of fluid-structure interaction of multiple scales has been developed. Each component of the system has been described by a reduced model (either 0D or 1D) and the different models have been integrated into the control system that governs the process.

Sketch of reduced FSI models coupled for two filling systems


Currently, other reduction strategies are being explored to obtain fast and reliable simulations of the complex dynamics of 3D fluids in some specific part of the system (such as the back pressure flange). For this purpose, a reduced order model based on the POD approach has been developed that will allow efficient parametric analysis in the design process.

Pressure field through the counter pressure flange: high fidelity FEM solution (left), reduced order POD solution (center) and absolute error (right)