Proper grain drying is a critical post-harvest process in the global agriculture industry. It is essential for preventing spoilage and ensuring long-term food security. The core challenge is to remove moisture from individual grains efficiently without damaging them. This involves complex Heat and Mass Transfer phenomena, where heat moves from hot air to the grains, and moisture moves from the grains into the air. Traditional drying methods are often inefficient, leading to high energy consumption and non-uniform drying.

To optimize this process, engineers are turning to advanced numerical modeling. A Drying Grains simulation offers a powerful way to investigate airflow, temperature distribution, and moisture removal within a grain bed. However, simulating this process presents a unique challenge: grains are a dense collection of discrete particles. To capture this physics accurately, we must use a coupled simulation approach. This project details a DDPM-DEM simulation in ANSYS Fluent. We use the Discrete Element Method (DEM) to model the particle-particle interactions and the Dense Discrete Phase Model (DDPM) to handle the fluid-particle coupling between thousands of grains and the air. For more on particle modeling, you can explore our DEM CFD simulation resources. This methodology is based on leading research, such as the work by Azmir et al. [1], which also discusses particle shrinkage.

• Reference [1]: Azmir, Jannatul, Qinfu Hou, and Aibing Yu. “CFD-DEM simulation of drying of food grains with particle shrinkage.” Powder Technology343 (2019): 792-802.

Figure 1- An example of traditional grain drying systems which modern CFD-DEM simulations aim to improve.

Simulation Process: CFD-DEM Coupling with Fluent and MATLAB

The simulation was built upon a simple 2D rectangular geometry created in ANSYS Design Modeler, with a structured mesh applied to accurately resolve the multiphase flow. The complexity of this Drying Grains CFD project lies entirely in the physics setup within ANSYS Fluent. Given the dense nature of the granular flow, an Eulerian-Lagrangian approach was selected.

Three critical models were combined to make this Drying Grains DDPM-DEM simulation possible. First, the Dense Discrete Phase Model (DDPM) was activated. This is essential because the volume fraction of grains is high, and DDPM accounts for the effect of particle volume on the continuous air phase. Second, to simulate physical interactions like collisions and friction between individual grains, the Discrete Element Method (DEM) was enabled. A standard DPM simulation would fail here because it cannot account for these crucial particle-particle interactions. Third, to track evaporated moisture, the Species Transport model was used to solve for water vapor in the air. A key innovation in this workflow was the MATLAB coupling. A custom MATLAB script was written to generate an initial particle position file. This file was used for the DPM Injection, ensuring the grains started in an organized array without unrealistic overlap, a critical step for a stable DDPM-DEM CFD simulation.

Post-processing: Analysis of Moisture and Temperature Dynamics

The simulation results provide clear insight into the drying process. Figure 2 shows the water mass fraction within the individual grains over time. The analysis of over 5,000 particles reveals a highly non-uniform drying pattern. The moisture mass ratios range from 0 up to 0.005319. The highest moisture content remains concentrated in the upper layers of the bed. This happens because the hot drying air enters from the bottom. As it travels upwards, it becomes saturated with vapor and loses its capacity to absorb more water. The simulation accurately captures the formation of “moisture channels,” which are preferential paths the air takes through the grain bed. This non-uniform behavior is a well-known phenomenon, and its successful prediction validates our DDPM-DEM model.

Figure 2: Water mass fraction distribution showing moisture content within the particles during the DEM-CFD drying simulation.

Figure 3, which displays the temperature distribution of the particles, completes the story. We observe a temperature range from a cool 298.2 K to 300.0 K. Crucially, the coolest spots in the bed correspond perfectly with the areas that have the highest moisture content. This is a direct consequence of the physics of evaporation: as water turns into vapor, it absorbs energy from the grain in a process known as the latent heat of vaporization. This actively cools the grain down. The small overall temperature difference (1.8 K) indicates that heat transfer from the air to the grains is relatively fast. However, the persistence of these cool, wet zones demonstrates that mass transfer—the physical removal of moisture—is the much slower, rate-limiting step. This fundamental insight, revealed by our Drying Grains CFD simulation, is exactly the kind of knowledge engineers need to design more efficient drying devices.

Figure 3: Particle temperature distribution across the grain bed, highlighting cooler zones where evaporation is most active.

Key Takeaways & FAQ

• Q: What is the difference between DPM, DEM, and DDPM in grain drying simulations?
  • A: DPM (Discrete Phase Model) tracks particles but ignores collisions between them, which is only valid for very dilute flows. DEM (Discrete Element Method) calculates the physical collisions and forces between every single particle, which is essential for dense beds of grains. DDPM (Dense Discrete Phase Model) is the method in ANSYS Fluent that allows us to combine DEM with the airflow simulation, accounting for the volume the particles occupy.
• Q: Why is MATLAB coupling used in this simulation?
  • A: In a Drying Grains simulation, defining the initial position of thousands of grains inside the dryer is difficult. If particles overlap at the start, the simulation will crash. A MATLAB coupling script is used to mathematically calculate a perfect, non-overlapping arrangement of coordinates, which is then imported into Fluent as an injection file.
• Q: How does the simulation model moisture removal?
  • A: The simulation uses the Species Transport model combined with a custom evaporation law. The solver calculates the difference in water concentration between the grain surface and the surrounding air. This gradient drives the mass transfer, converting liquid water in the grain into water vapor species in the air, while simultaneously removing latent heat from the particle.

The animation extracted from the transient (unsteady) CFD simulation is shown below: