An automated nozzle controller for self-propelled sprayers

摘要

Pesticide application is a vital, integrated component of 21st century agriculture. Pesticides allow more produce to be generated from fewer acres, increasing the world\u27s capacity and improving quality of life. Pesticide use however, is not independent of concerns. Pesticides by nature are destroyers. When applied to target pests, their destructive nature can be advantageously utilized, however when misapplication unites pesticides and susceptible non-target organisms, resulting effects can be catastrophic.The airborne movement of pesticides, spray drift, can result in up to 36.6% of the applied pesticide volume transporting outside of the intended swath to non-target organisms under high drift potential conditions (Grover et al., 1997). Studies have shown that through the implementation of best management principles, namely spraying with large droplet sizes, drift is reduced to less than 1% of the applied volume (SDTF, 1997; Grover et al., 1997). State-of-the-art drift reduction technologies inform applicators of real-time, site-specific dangers of drift, prompting applicators to implement best management practices. These technologies rely on the applicator for the decision making and implementation processes, adding subjectivity to the system and consequently, suboptimal performance. Objective, scientific decision making avenues are required for the future development of automated nozzle selection controllers to reduce spray drift.A basis for automated nozzle control was developed, implemented, and tested in the form of a tier 1 nozzle controller. Decision making processes rely on an on-board, real-time risk assessment; the comparison of mapped predicted depositions to established acceptable levels of depositions in sensitive areas. In-field testing results indicated the critical roles of a high-resolution representation of the nozzle spectrum (specifically for droplets \u3c 150 ym), and a regression model maintaining specificity within overall predictive accuracy. The nozzle controller was found to theoretically protect sensitive areas from excessive drift however significant differences between the predicted and actual drift phenomenon led to depositions measured in sensitive areas exceeding acceptable levels. Attempting to account for real-time operating conditions was found to significantly reduce the predictive accuracy of the controller, largely due to insufficient representation of highly variable wind speeds and direction vectors acting on droplets after release. Further development of predictive capabilities in representing wind speed and direction for durations up to 30 seconds after a droplet is released are required for micro-scale nozzle control.