Control Model and Experiment of Water and Fertilizer Proportion of Fertilizer Machine Based on RBF Neural Network

Water and fertilizer control equipment generally controls different matching ratio between water and fertilizer to achieve the integration of water and fertilizer irrigation. Most of the current machine for matching water and fertilizer cannot make fertilizer flow quickly track changes of water flow. To solve this problem, first of all, STM32F103ZET6 acts as the core of controller, the controller will output different PWM signal to change output voltage of DC motor driver, then the fertilizer flow can vary according to the water flow changes, and keep water and fertilizer holding at a set proportion; Secondly, in order to ensure the control precision, incremental PID algorithm is used to control the ratio of water and fertilizer, and RBF neural network is used to build the online parameter setting model to achieve precise and quick ratio control. Finally, using incremental PID and RBF- PID model, the simulation model is designed and different flow ratio and water flow are set up to validate the test. The result shows that the 5 kinds of different flow ratios and changes in water flow conditions, the highest average relative error of RBF-PID model is 7.55%, the highest coefficient of variation is 11.9%, the slowest flow ratio arrival time is 13s; through the test, RBF - PID model is more accurate and stable controlling than the incremental PID model.