Design optimization of rotary drum blade configuration for an e-powered mechanical weeding system using hybrid RSM–ANN models

Efficient weed management is essential for improving crop productivity and ensuring sustainable agricultural practices. The present study focused on the design optimization of weeding drum blade configuration of an e-powered mechanical weeder using response surface methodology (RSM) and artificial neural network (ANN). A face-centered central composite design (FCCCD) was employed with three independent variables namely number of blades (8–12), blade angle (20–30°), and blade height (25–45 mm). The performance responses evaluated were weeding efficiency and power consumption. The results revealed that weeding efficiency varied from 69.08% to 93.61%, while power consumption ranged from 146.93 to 247.58 W, indicating significant influence of blade configuration on system performance. The quadratic regression models developed using RSM were highly significant ( p < 0.0001), with coefficients of determination (R²) of 0.97 and 0.98, RMSE values of 1.30 and 3.39, and MAE values of 0.91 and 2.67 for weeding efficiency and power consumption, respectively. The optimization using RSM identified the optimal configuration of 10 blades, 20° blade angle, and 35 mm blade height, with an overall desirability of 0.77. This configuration provided a higher predicted weeding efficiency of 87.93 ± 1.27% with a minimum power consumption of 170.21 ± 3.32 W. The ANN model demonstrated superior predictive capability compared to the RSM model, with higher coefficient of determination (R²) of 0.98 for weeding efficiency and 0.99 for power consumption, along with lower prediction errors. The ANN model exhibited lower RMSE and MAE values of 0.89 and 0.71, respectively, for weeding efficiency and 2.98 and 2.14, respectively, for power consumption. The validated RSM–ANN modeling framework provides a reliable and efficient approach for optimizing blade configurations and enhancing the operational performance of e-powered mechanical weeding system.