PROJECT DETAILS

• Client: GRAHAM PACKAGING
• Date: 2023-2024
• Technology: DEM simulation

INTRODUCTION

In this post we will talk about a recent ICEMM project in which we use DEM (Discrete Element Method) simulation technology. This project is carried out with Hexagon’s Cradle ScFlow tool.

In this project we optimised a storage system for plastic food containers. The storage time of the containers is limited by the load that they support, since the deformation increases over time due to the creep phenomenon. The following image shows a typical time-deformation curve of this phenomenon:

In order to reduce the pressure in the containers, it is common to introduce mechanical elements in the silos that support part of the load. These solutions must meet minimum requirements in terms of constructive feasibility and maintainability. They must also reduce the storage capacity to a minimum and ensure a reasonable filling and emptying speed, without blockages that could penalise process times.

To study the movement and loads of individual containers with different construction solutions, DEM (Discrete Element Method) simulation technology is very useful. These simulations resolve the contact forces between the containers by iteratively updating their acceleration and position over time.

The distribution of loads in the bottles obtained from this simulation will allow the evaluation of the expected residence time, and the visualisation of the movement will give data on the admissible filling and emptying rates.

CALIBRATION OF PARTICLE MODELS

In order to obtain realistic data from DEM simulations, a fundamental step is the calibration of the properties of the numerical method.

In order to obtain realistic data from DEM simulations, a fundamental phase is the calibration of the properties of the method. First, a cluster of spheres is defined with a shape equivalent to that of the real container. This modelling method allows the computational advantages of the spherical particles to be maintained, but captures the behaviour of the containers in an appropriate way. A simplified example can be seen in the following image: numerical.

A series of tests are then carried out with real containers to calibrate the numerical parameters of the simulation. These include rebound tests on rigid surfaces and on other containers to obtain the “Coefficient of Restitution” and friction and rolling tests to calibrate these coefficients. Other data for the simulation, such as the mass of the containers or the stiffness of the material, were known from previous tests and measurements.

The following image shows a snapshot of the rollover and stacking simulation used to calibrate the coefficient of friction:

With this data calibrated to show realistic behaviour in the simulations, the study of storage systems can proceed.

VALIDATION OF DEM SIMULATION FOR THE CALCULATION OF CONTACT LOADS BETWEEN CONTAINERS

In this project we need to obtain load distributions on a set of large bottles. To display these results we decided to develop a ScFlow script to export the loads on each container and convert this data into a cumulative histogram of loads.

To validate that this method is representative, we started by simulating a series of stacked bottle columns without any load reduction system, to check that the evolution of the load distribution was as expected.

In this phase, other functions needed in the final simulation are also validated, for example the introduction of moving elements in the simulation and their interaction with the particles or the introduction of a small randomness in the particle generation to represent a more realistic behaviour.

STUDY OF CONFIGURATIONS WITH EXPERIMENTAL DATA

In the next stage of the project, we analysed DEM load reduction configurations already built by simulation. For these cases, experimental data were available for the dwell time of the cylinders with acceptable deformations.

These simulations allowed us to associate acceptable times to different load distributions, giving references against which to compare the histograms of the new proposals.

SYSTEM OPTIMISATION

In the last phase of the project we started working with the new proposed load reduction system, in which there were several geometrical parameters to be defined in order to achieve an optimal system.

The study methodology began by defining a series of configurations with constructive and maintenance feasibility. For these, filling and emptying is simulated with a simplified bottle model. Afterwards, the load distribution is simulated at maximum storage capacity with a detailed model. Based on the results, the next batch of configurations to be simulated is defined until a configuration that meets the customer’s requirements is reached.

This DEM simulation process makes it possible to evaluate the quality of a construction solution in terms of load reduction and filling and emptying rates. In addition, the geometry of the solution can be used to evaluate the constructive feasibility, maintainability and reduction of storage capacity. This allows an overall technical and economical assessment of the solution to be made prior to construction.