# Study on the variable side - deep fertilization control system for rice based on soil electrical conductivity

## TL;DR

This study develops a smart fertilization system for rice that adjusts fertilizer rates based on real-time soil conditions, improving efficiency and reducing waste.

## Contribution

A novel variable-rate fertilization control system for rice using real-time soil electrical conductivity (EC) detection and adaptive algorithms.

## Key findings

- The system achieved an average relative error of 2.70% in soil EC detection.
- The fertilization system had a maximum response time of 1.60 seconds and reduced fertilizer use by an average of 12.39%.

## Abstract

To address the problem that current rice fertilization devices rely on experience-based settings of fertilizer application rates and are unable to dynamically adjust according to soil fertility, resulting in low fertilizer use efficiency, a rice side-deep variable-rate fertilization control system based on real-time soil electrical conductivity (EC) detection was designed. First, a three-factor, four-level full factorial experiment was conducted to investigate the effects of soil moisture content, electrode insertion depth, and soil temperature on soil EC, and an EC calibration model was established based on an RBF neural network. Second, a fertilization strategy was developed and a fertilizer application model was constructed based on real-time EC, target yield, and implement forward speed; meanwhile, an incremental PID algorithm was adopted to achieve closed-loop control of variable-rate fertilization with motor speed as the control objective, completing the control system design. Finally, the system was deployed on a pneumatic groove-wheel fertilizer metering device and field experiments were carried out. The performance of the soil EC detection system and the variable-rate fertilization system was validated through EC detection accuracy tests and variable-rate fertilization system response tests. The results showed that the average relative error of EC was 2.70%, the maximum coefficient of variation of the fertilization system response stability was 3.98%, the maximum response time was 1.60 s and the average was 1.28 s, and the average fertilizer reduction rate was 12.39%. These results indicate that the proposed soil EC detection and variable-rate fertilization system can achieve rapid and accurate variable-rate fertilization operations. This study can provide equipment and technical support for rice side-deep variable-rate fertilization.