Sensor and control for consistent seed drill coulter depth

• Published in: Computers and electronics in agriculture 2016-09, Vol.127, p.690-698
• Main Authors: Nielsen, Søren Kirkegaard, Nørremark, Michael, Green, Ole
• Format: Article
• Language: English
• Subjects: Angle sensor; Control systems; Coulter control systems; Deviation; Dynamical systems; Sand; Sand and gravel; Seed drill coulter; Seed placement; Sensors; Soil (material); Soil resistance
• ISSN: 0168-1699; 1872-7107
• DOI: 10.1016/j.compag.2016.07.029

•The novel position system detected the high-frequency drill coulter depth vibrations.•By coulter down pressure control, the low-frequency depth variations were minimised.•The system provided a mean depth deviation from the desired coulter depth of ±1.2mm.•A three-position control system was found to be the best, cost-efficient solution. The consistent depth positioning of seeds is vital for achieving the optimum yield of agricultural crops. In state-of-the-art seeding machines, the depth of drill coulters will vary with changes in soil resistance. This paper presents the retrofitting of an angle sensor to the pivoting point of a drill coulter, providing sensor feedback to a control system that via an electro-hydraulic actuator delivers a constant coulter depth. The results showed a strong correlation between the angle of the coulter and the coulter depth under static (R2=1.00) and dynamic (R2=0.99) operations, verified by a sub-millimetre accurate positioning system (iGPS, Nikon Metrology NV, Belgium) mounted on the drill coulter. At a drill coulter depth of 55mm and controlled by an ordinary fixed spring loaded down force, the change in soil resistance reduced the mean depth by 23mm. By dynamically controlling the spring loaded down force based on the angle sensor, the mean depth was independent of the seedbed resistance change as shown from tests in soils ranging from sand to gravel. The PID controller was most effective because it providing a mean depth deviation from the target depth of −0.17mm and +0.08mm for sand and gravel, respectively. The most cost efficient control function was found to be the three-position control system, resulting in a mean depth deviation from the target depth of −0.89mm and −1.18mm for sand and gravel, respectively. A Fast Fourier Transform (FFT) analysis of the coulter depth measurements showed that the control system also provided a damping effect on the coulter depth variations. The research showed that it is possible to minimise the low-frequency drill coulter depth variations and provide a consistent coulter depth independent of soil conditions by using the developed sensor system and control system.