Phenomena: Heat transfer of mixing particles in gas
Potential Industry Applications:
- Food & Beverage
- Manufacturing
- Chemical
Main Software: STAR-CCM+
Analysis Type: CFD
Models:
-
Air flow: Turbulent k-ε model, gravity, ideal gas
-
Coupled energy transfer between particles and gas
-
Discrete Element Model (DEM): simulates each particle separately, assumes particles have constant properties (density), and are composite particles
Mesh: Rigid body rotational mesh
Solver: PISO solver
Findings:
-
Temperature profile simulated
-
Larger mass flow rate has higher average temperature but doesn’t affect the distribution
-
Simple mixing blades work best
-
Temperature and time most important quality
Petroncini develops coffee roasting machines that transform raw green beans into rich, full bodied roasts. In an effort to improve control and quality, Petroncini turned to CFD specialists Professor Luca Montorsi of the University of Modena and Reggio Emilia to simulate roasting half a ton of coffee in 15 minutes.
Roaster geometry.
The heart of any coffee production plant is the roasters, so that was the focus on Prof. Montorsi’s analysis, “Beans flow into the system and enter the roasting drum. There the beans encounter a rotor with blades that mixes the coffee beans inside the volume. The hot air escapes through the top.”
The simulation includes complex interactions between two models that simulate both the airflow and the beans.
The airflow is simulated using a turbulent k-ε model. The model takes into consideration coupled energy exchange between air and the beans. Gravity and ideal gas law were used in this model.
The beans, however, are simulated using a discrete element model (DEM). This model will simulate each bean’s movement inside the roaster separately. The model assumes the beans have constant properties (density), and that they are composite particles.
“We need to mesh the system for motion since the blades are rotating in the drum. So we chose a rigid body rotational mesh… The system itself was solved using a first order PISO solver,” said Montorsi.
The boundary conditions were as follows:
| Boundary | Value |
|---|---|
| Drum Rotation |
50 rpm |
| Inlet air temperature | 370 ͦC (700 ͦF) |
| Inlet coffee temperature | Room temperature |
| Mass (coffee) | 150 kg (330 lbs) |
| Particles (coffee beans) | ~1,500,000 beans |
| Air Inlet 1 | 3,500 Nm3/h (~2,000 cfm) |
| Air Inlet 2 | 2,500 Nm3/h (~1,500 cfm) |
“The temperature profile of the coffee bean is the variable most responsible for the quality of the coffee at the end of the roasting process. Simulating this curve, over time, is essential for the quality of the product,” noted Montorsi.
Comparison of average temperature and RMS temperature.
He added, “We can see that the larger mass flow rate enables higher average temperature, even though it doesn’t affect the temperature distribution as seen by the RMS value. Obviously, the larger mass flow rate enables a faster roasting process. Additionally, the global results show that the particle temperature is quite uniform in the roasting drums.”
Uniform temperature distribution in the roasting drum 3,500Nm3/h.
Uniform temperature distribution in the roasting drum 2,500Nm3/h.
The simulation also shows that the mixing of the beans is more affected by the complexity of the blade shape and rotor speed than the mass flow rate. Montorsi mentioned that, “the difference in bean velocity is of significance. The simple blade has the best results as the average value of the velocity is 20% higher than with the other blades. This is likely due to the fact that we have baffles in the middle of the complicated rotor which prevent the beans from falling through the rotor.”
The simulation was confirmed with the use of transparent pilot roasters. These pilot facilities use thermocouple sensors and high-speed cameras to determine the distributions. This allows the researchers to get an idea of the motion and physics on a pilot scale and verify their simulated results.
“Overall, our results show that temperature and time are the most important parameters for the quality of the product. You might end up with the same quality with a shorter time with higher temperature, or a low temperature for a longer time,” said Montorsi.
“Different blends require different parameters,” he adds. “For instance if we start with green coffee we have high moisture which will affect the temperature profile. Additionally, coffee drinkers in different countries prefer a different taste which in turn requires a different temperature and roast time. This is why simulation is important. It can address the temperature profile, quality, and influence of all the parameters.”
“As an Italian, this is a very important topic to me. But, I’m not going to tell you the best way to brew your coffee since everyone has their own recipe,” joked Montorsi. “But my favourite is espresso. In fact, another operation in the company is to design a capsule machine (@spresso) to make coffee. One secret of the coffee is the foam. For espresso the taste of the coffee is in the foam. We ensure proper foam is made using CFD multiphase simulations to develop the nozzle for the foam.” But perhaps that is an article for another time.