Optimizing Grain Milling via Python Simulation and Experimental Validation: Effect of Technical Clearance and Grain Physics
Description
• Integrated a Python-based simulation with experimental validation to optimize hammer mill performance for corn and barley. • Identified a 4 mm technical clearance as the universal optimum for maximizing the 500–1000 µm target particle fraction. • Established a high-fidelity correlation (R2 = 0.96) between physics-based simulation models and empirical milling data. • Demonstrated that grain mechanical resistance (e.g., 240.5 N for barley) significantly dictates clearance sensitivity and energy efficiency. • Provided a robust engineering tool for precision milling automation and reducing specific energy consumption in feed production.
Files
Steps to reproduce
1. The required physical and mechanical properties were measured after consulting artificial intelligence to provide the essential data needed for simulating the mill. Data for two types of grains (with different properties and resistance to breakage), which could not be performed experimentally, were obtained from scientific literature, such as corn and barley. 2. I was provided with Python code to conduct the simulation in the user-friendly Google Collab environment. The simulation was performed, the results were obtained, and the code was saved. 3. Two laboratory experiments were conducted using a commercially available mill commonly used in the region. Its design details, including rotor speed, power, and mill specifications and dimensions, were provided in the simulation code before the practical experiments. Two sieve opening levels and three clearance levels were used. The clearance was controlled by creating grooves to stabilize the edges of the circular milling machine. Three consecutive grooves were created at intervals that determined the varying clearances, and each treatment was repeated three times. 4. The results were obtained and analyzed using an ANOVA table. Figures were plotted using expert design software for statistical accuracy and to determine the optimal point for productivity, specific energy consumption, and the target grinding ratio for particle sizes from 500 to 1000 microns for chicken feed production. 5. Multiple regression equations were obtained for each grain removal method, and the predicted values for the target carrot yield were derived from these equations. 6. The prediction results from the simulations using physical equations were compared with the experimental results and the results from the regression equations. This comparison was conducted to verify the accuracy of the simulation code, which proved to be highly accurate and encouraging for use by small and medium-sized producers and manufacturers. Recommendations were made for mill manufacturers. 7. We provided user interface code for grain type selection. The program offers optimal settings, recognizing the grain type and adjusting based on differences in material resistance.
Institutions
- University of Basrah College of Agriculture