Variable rate fertilization control system for liquid fertilizer based on genetic algorithm

Variable rate fertilization has generally been implemented in the field traction liquid fertilizer applicator under a variety of soil and tillage conditions in recent years. However, it is very necessary to improve the precision, even fertilization, and fertilizer saving during operation in modern mechanized agriculture. In this research, a novel fuzzy PID control was proposed using genetic algorithm (GA) for the variable rate fertilization control system of liquid fertilizer. Firstly, a closed-loop negative feedback model was established for the control system of liquid fertilizer variable rate fertilization, thereby obtaining the transfer function of control process. The control process was also optimized, according to the structure of traction variable liquid fertilizer applicator and the composition of electrical components. Among them, the control object was mainly an electric proportional valve in the control system. The feedback channel was read by the flow meter and then transferred the electric signal to the controller. Specifically, the controller was implemented to compare the flow reading with the vehicle speed and the amount of fertilizer required for the current field. The obtained data was converted into the control signal and then output to the electric proportional valve, so as to realize the negative feedback control of system. Some models were established for the traditional, fuzzy, and GA-based fuzzy PID control, according to the requirements of control system. Particularly, the fuzzy PID control model was first established before the GA-based fuzzy PID control model. The input quantity of fuzzy controller was set as the error and the error rate of change, while the output quantity was set as the compensation value of three parameters in the PID controller, where each input and output quantity was set to 7 fuzzy language values. Therefore, there were 49 fuzzy control rules in total. Subsequently, the fuzzy control rules were chromosome-coded within GA. The chromosomes of fuzzy control rules were then simulated and optimized to obtain the optimal fuzzy control rule table using genetic operators, such as selection, crossover, mutation. Correspondingly, the fuzzy PID controller was further set, according to the optimal fuzzy control rules. Finally, MATLAB software was also selected to simulate the traditional, fuzzy, and GA-based fuzzy PID control. Consequently, the response time of electric proportional valve control was 4.86 s for the variable rate fertilization control system using the GA-based fuzzy PID control, which was significantly shorter than the 8.4 s of the traditional PID control, and the fuzzy PID control of 7.32 s. An experimental platform was constructed to carry out the stability and variable control experiment for the flow control in the control system of liquid fertilizer variable rate fertilization. In addition, the flow error was measured during operation in the fertilization stability experiment. The average relative errors of control system were 5.19%, 3.40%, and 1.14%, respectively, corresponding to traditional PID control, fuzzy and GA-base fuzzy PID control during stable operation. The signal was collected and recorded for the actual vehicle speed change in the variable control experiment. The inputs were the collected vehicle speed signal to the controller through a signal generator, thereby measuring the flow when the vehicle speed changed. Consequently, the actual response times were 5.19, 4.12, and 3.21 s, respectively, corresponding to the three control modes. Additionally, the actual response time of GA-based and fuzzy PID reduced by 1.98 and 0.91s, compared with the traditional PID control. Anyway, the GA-based fuzzy PID control presented better response time to flow control in the variable rate fertilization control system than traditional and fuzzy PID control, indicating better operational stability.