This content is based on a thesis presented at the University of Engineering and Technology – UTEC, Peru, by me, Luiguie Sagastegui in partnership with Helard Álvarez. This thesis fills a gap in research in this area due to the lack of fundamental studies using state of the art first principle computational techniques.
The main purpose of the study is to develop a DEM-CFD coupled model to simulate a wheat cleaning aerodynamic machine.
Background & Motivation
There is a lack of fundamental studies due to the state of computational numerical studies and its applications in modern engineering, such as the case of coupled simulations between DEM and CFD (Discrete Element Method and Computational Dynamic Fluid, respectively).
These simulations explain the theory and the technical aspects of the aerodynamic cleaning process of a wheat cleaning machine. In addition, these numerical studies avoid traditional experimental testing and optimization, which previously were performed constantly to deal with the problems of low cleaning performance and over-demand of energy from the centrifugal fan motor of the machine.
According to literature,the causes of these problems are attributed to inefficient operational factors and design. Most common amongst these are poor calibration of the air-particle flows and the dimensions of the pneumatic conduit.
Through the aforementioned CFD-DEM coupled simulations, the operational and design factors can be freely varied in order to better understand the behavior of the particles under the influence of a forced air flow inside the wheat cleaning machine and achieve optimized parameters at a fraction of the cost and effort with conventional testing.
To better understand the behavior and wheat and brush particles, three 1 way DEM-CFD simulations were performed between Rocky DEM and ANSYS Fluent in order to simulate the process of cleaning the wheat-cleaning machine for three directions of airflow (0°, 22.5° and 45°).
The CFD simulation was started first and configured with an air flow velocity of 7.5 m/s (wheat terminal velocity) and a hydraulic diameter of the entrance area of 0.25 m for the three directions: Air flow velocities directed 0°, 22.5° and 45° with respect to the horizontal.
The flow field obtained from these CFD solutions were exported to Rocky DEM. In Rocky DEM, behavior of the wheat grains under the flow field was studies.
To begin the DEM simulations, physical and mechanical properties of the wheat and brush particles were configured (see Table 1: Properties required for the configuration of the particles in Rocky DEM.), the feed input ratio for both particles (see Table 2: Reason for entry of wheat and brush particles in Rocky DEM). And finally the unidirectional simulation was started with a time step of 0.05 seconds for a duration of 2 seconds for each of the three directions.
The results of the DEM-CFD simulations were obtained directly in Rocky DEM and were basically the monitoring of the translational velocity of wheat and brush particles in each time step for the 3 flows, which are shown in Figure 2.
In addition, the final displacement of the particles, obtained from the simulations, was validated and compared through experimental tests carried out in a test bench at scale configured under the same operation and design parameters, and a margin of 5% error for wheat and 10% for the chaff.
Following the results of the simulations carried out, the translational speed of wheat and the brush at the entrance and exit of the interaction zone were in the range of 0.5-1.5 m/s and 3-4.5 m/ s, respectively (see Figure 3), and for the three flows 0 °, 22.5 ° and 45 ° cleaning performances of 99%, 98% and 96% were given.
Additionally, the CFD configuration for the test bench is shown to scale (see Figure 4) which required 3 circular inputs to simulate the flows generated by the 3 hair dryers during the experimental tests. The results of the simulations of the test bench in Figure 5 are shown.
An experimentally validated 1 way CFD -DEM coupled simulation was set up to simulate the wheat cleaning processes and identify critical design and process parameters. The computational model predicted the real behavior of wheat and chaff particles with an accuracy of 5% and 10% respectively.
For more details or to read the complete study, you can download it on the banner below.
Luiguie Martin Sagastegui Ronceros
Mechanical Engineering student of the 10th cycle at the University of Engineering and Technology (UTEC) with a Major in Maintenance Management. First Mentor of the Materials Mechanics course at UTEC. Interested in developing in the maintenance area and in the area of innovation and project development.