Skip to main content
ARS Home » Southeast Area » Stoneville, Mississippi » Crop Production Systems Research » Research » Publications at this Location » Publication #245385
Title: Improving Flow Response of a Variable-rate Aerial Application System by Interactive Refinement

Author
item Thomson, Steven
item Huang, Yanbo
item Hanks, James
item Martin, Daniel - Dan

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/21/2010
Publication Date: 5/16/2010
Citation: Thomson, S.J., Huang, Y., Hanks, J.E., Martin, D.E. 2010. Improving Flow Response of a Variable-rate Aerial Application System by Interactive Refinement. Computers and Electronics in Agriculture. 73(1):99-104.

Interpretive Summary: Agricultural aircraft are used to apply chemical and biological pest-control agents, harvesting aids, and nutrients to assure a high-yielding and economically viable crop. Only recently have agricultural aircraft been equipped to implement variable-rate application to match site-specific needs of the crop. Variable-rate aerial application systems have seen only limited use since about 2004, and very little information has been presented on the accuracy of these systems for placement of chemical and response of these systems to changing rate requirements. Aerial flow control systems must also adjust flow properly to accommodate changes in ground speed. Previous studies conducted in our fields using an Air tractor 402B aircraft have quantified application error with respect to field boundaries whose precise locations were known. Spraying system response for a single group of spray runs was also illustrated, and responses tended to vary from overdamped to slightly underdamped depending on rate change levels. Potential improvements to the control algorithms were proposed. This study was designed to quantify system response to changing rates before and after the system manufacturer changed the control algorithm to improve response. Results for a set of south-north runs indicated reduction of average error from 6.9% before control algorithm modification to 1.8 % after algorithm modification. North-south runs indicated response lags after control algorithm modification and the system’s inability to meet flow demands at the highest flow rates. This was due to the system calling for more spray than the booms were able to deliver at high ground speeds. These factors biased post-modification results unfavorably when compared with results obtained before the control algorithms were modified. A custom-built flow monitor with constant integration interval of 0.1-s was also evaluated to improve resolution of data acquisition from the on-board flow meter. Benefits of the new flow monitor were illustrated by examining data from both 2005 and 2008 prescription runs. For the 2005 data, integration times per run matched expected values based on ground speed when using either the new flowmeter monitor or conventional monitoring via the AutoCal II with its irregular data integration intervals. The 2008 data showed inconsistencies in total integration time per run when reading flowmeter data via the Autocal II; these intervals varied between 1.2 and 1.66-s. Integration timing intervals matched expected results when using the new flowmeter monitor instead of the AutoCal II to read and output data. Inconsistencies in AutoCal II timing were attributed to possible modifications in loop-timing portions of the AutoCal II control algorithm since the 2005 run was conducted. These results further support the value of the new flowmeter monitor in providing consistent results regardless of changes the manufacturer might make to the AutoCal data acquisition and control program. The experiments illustrated how recursive refinement of control algorithms in collaboration with the control system manufacturer could improve system response characteristics. System evaluation techniques described herein should also be applicable to aircraft that use propeller-driven spray pumps as well as hydraulically controlled spray pumps.

Technical Abstract: Experiments were conducted to evaluate response of a variable-rate aerial application controller to changing flow rates and to improve its response at correspondingly varying system pressures. System improvements have been made by refinement of the control algorithms over time in collaboration with the system manufacturer, Houma Avionics, Houma, LA, USA. The variable-rate application system consists of Differential Global Positioning System (DGPS)-based guidance, AutoCal II automatic flow controller, and hydraulically controlled spray pump. The AutoCal II was evaluated for its ability to track desired flow rates set by the pilot. The system was then evaluated over several field trials to quantify its response to rapidly changing flow requirements and to determine the effect of the latest control algorithm improvements on response characteristics. System responses were analyzed while operating the AutoCal II in automatic mode over a pre-set field prescription containing four management zones (28, 47, 56, and 37 L/ha each 81 m long). To evaluate the effect of control algorithm improvements, areas under the flowrate-time curves were integrated and percentage differences in areas between those response curves and target flow rate curves were determined. Results for south-north runs indicated reduction of average error from 6.9% before control algorithm modification to 1.8 % after algorithm modification. A custom-built flow monitor with constant integration interval of 0.1-s was then used to improve resolution of data acquisition from the on-board flow meter. Benefits of the new flow monitor were illustrated by examining data from both 2005 and 2008 prescription runs. For the 2005 data, integration times per run matched expected values based on ground speed when using either the new flowmeter monitor or conventional monitoring via the AutoCal II with its irregular data integration intervals. The 2008 data showed inconsistencies in total integration time per run when reading flowmeter data via the Autocal II; these intervals varied between 1.2 and 1.66-s. Integration timing intervals matched expected results when using the new flowmeter monitor instead of the AutoCal II to read and output data. Inconsistencies in AutoCal II timing were attributed to possible modifications in loop-timing portions of the AutoCal II control algorithm since the 2005 run was conducted. These results further support the value of the new flowmeter monitor in providing consistent results regardless of changes the manufacturer might make to the AutoCal data acquisition and control program. The experiments served to illustrate an example of how recursive refinement of control algorithms in collaboration with the control system manufacturer could improve system response characteristics. System evaluation techniques described should also be applicable to aircraft that use propeller-driven spray pumps as well as hydraulically controlled spray pumps.