The purpose of this study is to investigate the effects of forward speed, soil moisture content, disc type and disc gang angle on soil disturbance area caused by a four-disc harrow in a soil bin. Moreover, the study aims to predict and optimize the soil disturbance area with considering the input variables to obtain the goals of conservation tillage. The input variables included forward speed (6.43 and 13.15 m/min), soil moisture content (5.37% and 13.25% (db)), disc type (plain and notched), and disc gang angle (0, 15 and 30°). Output variable was soil disturbance area which was measured using profile meter. A total of 72 tests were run based on randomized complete block design (RCBD). Test data were used to develop the regression, artificial neural network (ANN), and response surface methodology (RSM) models and their accuracy were compared. Furthermore, RSM approach was used for optimization process. Results indicated that soil disturbance area was significantly affected by disc gang angle (P < 0.01), the interaction between moisture and angle (P < 0.05), and speed and angle (P < 0.01). Increasing the angle from 0 to 15 and 30° substantially enhanced disturbance area (115.1, 222.1, and 337.7 cm², respectively). Regression, ANN and RSM models predicted the disturbance area of notched discs with higher accuracy (R2 = 0.8891, 0.9666 and 0.9696, respectively) than plain discs (R2 = 0.7747, 0.9356 and 0.8826, respectively). The optimization process for disturbance area were obtained 280.814 and 295.710 cm2 for equal moisture content of 9.31%, forward speed of 8.82 and 7.282 m/min and disc gang angle of 22.57 and 23.935° with equal desirability of 1.000 for plain and notched discs, respectively. Among the input variables under investigation, the angle of the plate group was the most effective on soil disturbance area. The best studied prediction approach was RSM for disturbance area of notched discs with highest R2 (0.9696). The obtained advantages of RSM approach relative to ANN and regression ones in this project are listed as higher accuracy in prediction of soil disturbance area, 3D surface curves in RSM output help readers to better understand the changes in the output parameter under the influence of the input parameters, they can also help researchers to obtain the output parameter in the unmeasured input parameters. Moreover, the optimization process could be done only using RSM. The RSM approach was able to obtain the optimal values of the input variables with a very high desirability (1.000) in order to idealize the soil disturbance rate.
Sheshpoli et al. (Mon,) studied this question.