Control system strategies for efficient liquid application of pulse width modulation (PWM) systems


Journal Title

Journal ISSN

Volume Title



Agricultural sprayers equipped with modern technologies such as pulse-width modulation (PWM) systems are widely used by farmers to efficiently use liquid application and mitigate environmental and human health concerns. These flow control systems can manage flow on a nozzle-by-nozzle basis and thus may provide uniform application pressure to maintain the uniform droplet size and target application rate. However, limited studies have been conducted which may help producers to make an informed decision in selecting a commercially available PWM system that can provide uniform spray coverage and droplet size. Therefore, two studies were conducted to 1) develop the control system strategies for the efficient liquid application for the PWM system and 2) evaluate the sensor-based variable rate nutrient application methods response to crop vigor and yield. A methodology was developed to quantify the spray coverage and droplet size uniformity of the self-propelled agricultural sprayers equipped with PWM technology during field operation. Two PWM systems, namely the John Deere ExactApply™ system (PWM1 system) and Case AIM Command Flex II™ system (PWM2 system), were operated at different combinations of frequency and duty cycle. The PWM1 system was operated 15Hz, and 30Hz, and the PWM2 system was operated at 10Hz, 15Hz, and 35%, 50%, and 75% duty cycle. To evaluate the performance of the PWM2 system at higher frequencies, the PWM2 system was operated at 10Hz, 15Hz, 20Hz, and 30Hz and at a 50% duty cycle. The results show that each PWM system provided different spray coverage uniformity when operated at different combinations of frequency and duty cycle. For the PWM1 system, the higher spray coverage, spray coverage uniformity, and droplet size uniformity was achieved when the system was operated at 30Hz regardless of the duty cycle settings. The PWM2 system provided higher spray coverage at 20Hz. However, the spray coverage uniformity, droplet size, and droplet size uniformity were significantly the same at 20Hz and 30Hz frequencies. Based on the first study's results, the PWM2 system was utilized to conduct sensor-based variable rate nutrient application. Two fields (Lafe's and Across from Kopfers) were selected for conducting field experiments, hereafter referred to as Field1 and Field2, respectively. Field1 had two strips (strip1 and strip2), and Field2 had a single strip (strip3) which were selected for the nutrient application. All the strips were planted with different hybrids of corn. Both fields were applied with nutrients equivalent to 100 pounds (lbs) more nitrogen (N) than the strips before the nutrient application. The other 100lbs of N equivalent nutrient (UAN32) was applied to the strips at the V6 stage. Two nutrient application methods, namely the constant rate (C.R.) and sensor-based variable rate application (VRA) were used on adjacent 18.29-mwide passes to apply the nutrient onto the strips. The Augmenta sensor and the Y-drops were mounted on the PWM system to perform the variable rate side-dress nutrient application. The application rate was set at 263.41Lha-1 (28.16 GPA) to apply the 100 lbs of N using the UAN32 fertilizer. The tissue analysis was performed to measure the total %N of the crop within each small strip at the V6 stage and silking (R1) stage using the dry combustion method. The aerial imagery was taken using the Micasense Rededge camera at the R1 stage to generate the NDRE maps to analyze the crop vigor of the strips. The results showed at the V6 stage, all the strips and the control contained the total %N within the range of 3.12-3.67%, and at the R1 stage, the total %N was significantly the same for C.R. and VRA but was lower for the control. Sensor-based VRA provided greater crop vigor (NDRE: 0.6239, 0.6535, 0.6584) when compared to the constant rate (NDRE: 0.6142, 0.6287, 0.6402) for strip1, strip2, and strip3, respectively. The crop yield for the VRA was 202.84 and 189.33 bushels per acre (bu.acˉ¹), which was significantly higher than the C.R. application having crop yield of 196.17 and 182.14 bu.acˉ¹ (p-value:0.0145 & 0.0146) for strip1 and strip2, respectively. For strip3, the crop yield was significantly the same for the VRA (199.74 bu.acˉ¹) and C.R. application (194.0 bu.acˉ¹). In conclusion, crop producers can operate the system at higher frequencies to achieve better spray coverage and droplet size uniformity. The producers could use sensor-based variable rate nutrient application to achieve greater crop vigor and yield. However, the operators should understand the PWM system components and the control system strategies for properly utilizing the technology to reduce off-target liquid application and improve crop yield and vigor. Future research and development may include the actual chemical spraying on the crops to validate the uniformity of spray coverage and droplet size by quantifying the effectiveness of weed/pest control. Additional future studies also need to be conducted to quantify real-time nutrient flow rates on the nozzle-by-nozzle control for the sensor-based variable rate nutrient application.



Pulse width modulation system, Volume mean diameter, Spray coverage and droplet uniformity, Variable rate application, Sensor-based nutrient application, Sprayers

Graduation Month



Master of Science


Department of Biological & Agricultural Engineering

Major Professor

Ajay Sharda