**3. Analysis of the Influence of the Rolling Friction Coefficient**

This section analyses the influence of the RFCP-P on the angle of repose by means of a repose angle test and the influence of the RFCP-B on the percentage passing by means of a self-flow screening test. The results confirmed that accurate calibration of the above two parameters is necessary.

#### *3.1. Repose Angle Simulation*

The RFCP-B was set to 0.025 and the RFCP-P was set to nine values of 0, 0.025, 0.05, 0.075, 0.1, 0.125, 0.15, 0.175, and 0.2. The single factor test was conducted to analyze the effect of the RFCP-P on the angle of repose.

The test procedures were same as in previous studies; the dimensions of the device were 220\*48 mm [19]. The EDEM (Version, 2018, School of Biological and Agricultural Engineering, Jilin University, Changchun, Jilin, China) software with the parameters in Table 1 was used to simulate the repose angle test. First, 4000 particles were generated in the particle factory. After two seconds, the insert plate was pulled out and the seed particles flowed out of the loading box; at the same time, the angle of repose was formed in the loading box. The repose angle results were analyzed using image processing software, as shown in Figure 9. The simulation tests were repeated five times for SN42, JD17, and ZD39, respectively, according to the above steps.

**Figure 9.** Image of (**a**) repose angle simulation results and (**b**) processing diagram.

#### *3.2. Self-Flow Screening Simulation*

The RFCP-P was set to 0.025 and the RFCP-B was set to nine level values of 0, 0.025, 0.05, 0.075, 0.1, 0.125, 0.15, 0.175, and 0.2, respectively. A single factor test was conducted to analyze the effect of the RFCP-B on the percentage passing.

The test procedures were the same as in previous studies [19].The inclination angle and aperture sizes of the three varieties of self-flow screening simulation were 11◦ and 8 mm, respectively. The EDEM software with the parameters in Table 1 was used to simulate the self-flow screening test. In the simulation, 1000 particles were generated in the particle factory. After the particles were stable, we pulled out the insert plate, and the soybean particles moved downward along the sieve. At the end of the movement, the particles were present in the receiving area below or on the sieve deck. The numbers of soybean particles in the corresponding areas were counted and numbered as 1–5, as shown in Figure 10.

**Figure 10.** The statistical areas of self-flow screening simulation.

#### *3.3. Analysis of Simulation Results*

The relationship between the angle of repose and RFCP-P is shown in Figure 11. With the increase in RFCP-P from 0 to 0.2, the increasing trend of the angle of repose of the three varieties was obvious. For SN42, the angle of repose gradually increased from 19.33 to 37.57◦; for JD17, the angle of repose gradually increased from 24.8 to 37.02◦; and for ZD39, the angle of repose gradually increased from 27.31 to 39.1◦. Analysis of the results shows that the RFCP-P had a significant effect on the angle of repose. Therefore, an accurate RFCP-P needs to be calibrated through calibration tests.

**Figure 11.** The relationship between angle of repose and RFCP-P for (**a**) SN42, (**b**) JD17, and (**c**) ZD39.

The relationship between percentage passing and RFCP-B is shown in Figure 12. For SN42, due to its high sphericity, the percentage passing did not change much when the RFCP-B gradually increased from 0 to 0.075, whereas the percentage passing tended to significantly decrease when the RFCP-B gradually increased from 0.075 to 0.2, with a range of 94.53 to 29.23%. For JD17 and ZD39, the percentage passing was significantly reduced as the RFCP-B increased, with a range of 95.43 to 26.4% and 86.97 to 23.53%, respectively. The analysis showed that the RFCP-B had a significant effect on the percentage passing. Therefore, an accurate RFCP-B needs to be calibrated by calibration tests.

**Figure 12.** The relationship between percentage passing and RFCP-B for (**a**) SN42, (**b**) JD17, and (**c**) ZD39.

### **4. Study on the Sensitivity of RFCP-P and RFCP-B**

#### *4.1. Comprehensive Simulation Test of Sensitivity Analysis*

As mentioned before, the RFCP-P and RFCP-B need to be calibrated. If two parameters are simultaneously calibrated, there are multiple sets of solutions. Therefore, the sensitivity of RFCP-P and RFCP-B to the angle of repose was first analyzed by simulating pitch angle tests. Seven levels of 0.01–0.07 were taken for both RFCP-P and RFCP-B, and a full 7\*7 simulation was performed.

#### *4.2. Analysis of the Results*

The results of the effect of RFCP-P and RFCP-B on the angle of repose for the three varieties of soybean seed particles are shown in Figure 13. For the three varieties, the effect of RFCP-P on the angle of repose was highly significant, whereas the effect of RFCP-B on the angle of repose was insignificant.

**Figure 13.** *Cont*.

**Figure 13.** Sensitivity analysis of RFCP-P and RFCP-B to angle of repose for (**a**) SN42, (**b**) JD17, and (**c**) ZD39.

According to the above analysis, the RFCP-P can be calibrated by a single factor test of the repose angle.
