Pneumatic conveying is widely used to transport bulk solids in industrial manufacturing processes because it offers a safe, flexible, and automation-friendly method of material handling. Even so, systems of this kind can still face issues such as plug flow, agglomeration, and blockages. These challenges often reduce process control, increase maintenance demands, and negatively affect product quality. Fine bulk solids are particularly prone to these problems. In many cases, the root cause can be traced to the hopper feeder section of the pneumatic line, where arching and ratholing may lead to intermittent discharge and pressure surges.
A clearer understanding of the interaction between operating conditions and bulk material properties can help improve process performance and reliability. Physical testing remains important, but relying on experiments alone can be expensive and time-consuming. Numerical modelling therefore becomes a valuable part of the investigation, helping reduce development costs while also providing deeper insight into system behaviour.
Numerical modelling and analysis of multiphase systems such as pneumatic conveying lines can be carried out using a two-way coupled CFD-DEM approach. In this workflow, the motion of the bulk solid is resolved in Simcenter EDEM software, while the airflow is resolved in Simcenter Acusolve software. This coupled methodology makes it possible to study the interaction between the solid and fluid phases in far greater detail.
The modelling of a pneumatic hopper feeder system was carried out in collaboration with Novo Nordisk, with the aim of linking system performance to pressure drop and powder flowability. The system consists of an axisymmetric hopper connected to a horizontal pipe, with an air valve positioned downstream of the hopper, as shown in Figure 1.
Figure 1. A two-way coupled CFD-DEM model of a pneumatic hopper feeder system
Within Simcenter EDEM software, the bulk solid is represented using a meso-scale modelling approach. In this method, numerical particles are defined at an intermediate scale between the actual particle size and the full system scale so that practical computational times can be achieved. The influence of particle shape is captured using a multi-sphere approach, while the Edinburgh Elastic-Plastic-Adhesive contact model is applied to represent the visco-elastic-plastic-cohesive behaviour of the powder.
The bulk solid model is selected automatically from the EDEM powders database using direct shear test measurements for micro-crystalline cellulose as the search criteria. In addition, the particle-to-wall interaction parameters for the hopper are calibrated against wall yield locus measurements of micro-crystalline cellulose sliding against stainless steel.
For the fluid-solid coupling, an Euler-Lagrange methodology is used. The fluid domain is discretised in Simcenter Acusolve software using an unstructured tetrahedral mesh. An unresolved CFD-DEM bi-directional coupling strategy is then applied to account for the influence of particle motion on the fluid flow. Under this formulation, the fluid phase is governed through the volume-averaged Navier-Stokes equations, based on the local averaging of fluid-particle interaction forces and solid volume fraction. The drag force acting on the particles is calculated using the Gidaspow correlation.
The predictive capability of the numerical model is validated by comparing the simulation results with experimental measurements of the mass flow rate of micro-crystalline cellulose in the pneumatic hopper system.
Validation cases are considered for both a fully closed air valve and a fully open air valve so that the pressure drop extremes of the system can be covered. In the closed-valve case, the air opening is completely blocked and the outlet pressure is varied. In the open-valve case, a constant airflow rate, obtained from experimental measurements, is applied at the air valve inlet. A comparison of the simulated and experimental results is shown in Table 1, while the computed air velocity and particle volume fraction fields for the closed-valve and open-valve cases are shown in Figures 2 and 3.
A strong quantitative agreement is observed between the predicted and measured mass flow rates for both cases. In addition, the simulated bulk solids flow patterns qualitatively align with the experimental observations. Taken together, these findings support the validity of the modelling approach.
| Valve configuration | Pressure drop (kPa) | Experiment (kg) | Simcenter Acusolve-Simcenter EDEM (kg) |
|---|---|---|---|
| Closed | 20 | 2.126 | 2.538 |
| Closed | 22 | 3.303 | 2.727 |
| Closed | 23 | 3.088 | 2.645 |
| Open | 2 | 0.466 | 0.375 |
Table 1. Comparison of the transferred powder mass between simulation and experiment
Figure 2. Velocity magnitude and solid volume fraction contour plot for the closed-valve case at 23 kPa pressure drop
Figure 3. Velocity magnitude and solid volume fraction contour plot for the open-valve case at 2 kPa pressure drop
The numerical model for the closed-valve case is then used to perform a parametric study aimed at relating applied pressure drop and bulk solids flowability to process performance.
To investigate this further, the discharge behaviour of two flowability extremes is simulated across a wide pressure-drop range. The predicted evolution of mass flow rate for the lower-cohesion material at varying pressure drops is shown in Figure 4, while Figure 5 illustrates the effect of cohesion on discharge behaviour at a fixed pressure drop.
Figure 4. Predicted mass flow rate at varying pressure drops for a material of low cohesion
Figure 5. Predicted mass flow rate at 20 kPa pressure drop for materials of two cohesion extremes
The results reveal a transitional flow regime characterised by a shift from continuous discharge to intermittent arching behaviour, as illustrated in Figure 6. This response is driven by the complex interaction between outlet pressure and powder flowability. As the pressure drop increases, the consolidating stresses within the powder bed also rise. This, in turn, increases the bulk material strength in the arch region and raises the likelihood of arch formation.
Because of this intermittent arching behaviour, the relationship between mass flow rate and pressure drop becomes non-linear. Rather than increasing proportionally, the mass flow rate rises sub-linearly with increasing pressure drop.
Bulk cohesion has an even stronger effect on arch stability. More cohesive materials show highly discontinuous flow behaviour, even when discharging under relatively low pressure drops. Following the collapse of an arch, a significant localised increase in consolidating stress can also be seen near the outlet region, as shown in Figure 6. This local stress build-up may contribute to unwanted agglomeration and material accumulation. These findings are consistent with experimental observations of material build-up in the region of the pipe elbow.
Figure 6. Intermittent arching behaviour during discharge of a low-flowability bulk solid
The study suggests that simply increasing pressure drop is not always an effective solution for poor flow in a pneumatic hopper feeder system. Instead, the relationship between pressure drop, consolidating stress, and powder flowability should be considered carefully when evaluating and improving discharge performance.
Looking to simulate complex bulk material and fluid interactions with greater confidence? Request a quote for Simcenter EDEM and Simcenter Acusolve to explore the right solution for your modelling and process requirements.
Content adapted from a blog published on Altair Community.
