Bulk Engineering, Bulk Technology, Modelling

Bulk solids handling design using DEM modelling

Enes Kaya, senior project engineer at Jenike & Johanson, explains how advances in computational modelling techniques have improved bulk solids handling design.

Enes Kaya, senior project engineer at Jenike & Johanson, explains how advances in computational modelling techniques have improved bulk solids handling design.

In the development of bulk material handling equipment, engineers have traditionally relied on empirical formulas, physical scale modelling, and past experiences as their primary design tools. While these methods have been refined over time, with varying levels of industry success, each has their limitations. Fortunately, advances in computing power have introduced several computational modelling techniques, such as the Discrete Element Method (DEM). 

DEM allows engineers to model the mechanics of individual particles to simulate bulk material behaviour. By applying Newton’s Laws of Motion, we can calculate the forces, accelerations, velocities, and positions of each particle, which in turn, predicts the overall behaviour of the bulk material. 

It is important to note that DEM simulations are computationally intensive, especially for systems with large number of particles. To ensure stability and efficiency, an appropriate time step is needed at which each particle force and displacement calculation is performed. It requires a significant amount of time to run these calculations. Incorporating DEM into the design workflow can provide valuable insights for improving equipment design by providing a visual material rheology where sensors and field measurements cannot.

With DEM, a granular bulk solid is treated as a system of interacting particles. Each particle interaction follows a soft contact approach where an overlap of particles is accepted. These particles are treated as rigid bodies, typically of varying size and shape, and the interactions between them governed by a contact model, energy dissipation by friction, and inelastic collisions.

Related stories:

The choice of the contact model and the determination of the material properties are essential for calibration and dependent on the application studied. For example, to evaluate the flow-ability of cohesive materials under high storage stresses for systems that may experience blockages such as ratholing or arching, the stress history of the material needs to be considered. The DEM parameters may differ for the same material in a dilute flow system where material is not confined and in constant motion.

In its initial stages, DEM used 2D disks and later in 3D, spheres. The benefit of this shape is a simplification of the contact detection model which vastly reduces the computational power required. With the increased computational performance achieved from today’s CPUs and GPUs, particles of varying shapes can be simulated. Advanced software packages can now take point cloud scan data to create and import custom shapes into the simulation. However, with increased complexity of the particle geometry, the required computational time can also be prohibitive to a simulation.

How DEM works in practical applications

A DEM simulation will need the boundary walls of the system to be defined, such as a hopper, chute, mixer etc., as well as the motion of moving components.

The typical flowchart of a DEM program.

The second step is determining the suitable particle size and shape representative of the bulk material being studied. A calibration is performed under a contact model characteristic of application behaviour. Depending on the selected contact model, the physical properties of the material such as the particle density, frictional parameters, modulus of elasticity, Poisson’s ratio, coefficient of restitution is set. While these parameters may more easily be determined and measured for large, free flowing materials; fine and cohesive materials require additional tests to validate the model.

Limitations of DEM

Although DEM can provide insight in the design of numerous bulk material handling systems as an added tool to study material flow behaviour, some limitations exist. One of the primary limitations of DEM simulations is their computational intensity. The simulation run time increases proportionally with the number of particles simulated in the system. This can make it challenging to accurately simulate complex systems requiring large number of particles and vast ranges of particle sizes. Additionally, the accuracy of the selected contact model in simulating the desired behaviour should be considered in determining how representative the results are.

Reducing computational burden

The run time of a simulation could be reduced by: 

  • Increasing the number of GPU’s and CPU’s in the simulation
  • Reducing particle count by only simulating the regions of interest.
  • Reducing the complexity of the system, by simplifying the geometries and only simulating internal surfaces. Don’t simulate the boundary surfaces of bolts, nuts, flanges that join geometries.
  • Model spherical particles where feasible 

The contact model has negligible effect on the total run time of the simulation, but it has a significant effect on the accuracy of the simulation results. To ensure the that the contact force model accurately represents the physical system reality, the user must fully understand the model and how modifications to various parameters will affect the outcome of the simulation.

Conclusion

DEM has become a very useful tool for designers and engineers to predict the behaviour of bulk materials in a system. There are key considerations such as setting the correct boundary conditions, selecting the most suitable contact model, and determining the correct physical properties of the materials to accurately simulate the system. The accuracy of DEM simulations rely on the users understanding of not only the physical properties of the bulk materials, but also their understanding of the system as a whole. 

Send this to a friend