| Spatially-variable control of fertilizer and pesticide application has a great potential to increase benefit from crop yield and significantly reduce the introduction of fertilizers and pesticides into the environment. However, inaccuracies in application rate control and application errors caused by system dynamics, nonuniform distribution of flow characteristics, and mixture transportation delay will reduce that potential.; In this research, state-space models of selected liquid spraying systems were derived. Optimal pump flow rate control using Linear Quadratic Gaussian (LQG) control theory was conducted on a dual-pump spraying system. Pump flow rate estimation using a pressure transducer has been implemented with a Kalman filter.; A Geo/SQL GIS based field data handling system was built to automatically determine application rates for spatially-variable controlled liquid fertilization. Field data from a German field was used to build the spatial database.; Tools for analysis and design of spraying systems were developed. The flow characteristics along a boom hose has been studied as a divided flow manifold problem. Application error caused by system transient response, nonuniform flow distribution, and transportation delay was formulated. Optimal boom hose diameter and precommand time concepts based on minimizing integral absolute application error were derived and examples were demonstrated. Programs for flow characteristic and optimal boom hose diameter and precommand time selection were developed.; Overall this study supports the development of engineering techniques for designing spatially-variable liquid fertilizer/pesticide application. |
