基于气流辅助的颗粒肥流量光电检测系统设计与试验

Design and Experiment on Photoelectric Detection System for Flow Rate of Granular Fertilizer Based on Airflow Assistance

摘要:

针对颗粒肥料相互遮挡以及外槽轮排肥器引起的肥料分布不均现象，造成颗粒肥流量难以高精度检测的问题，本文提出一种基于气流辅助的颗粒肥流量光电检测方法，设计一种气流辅助光电式颗粒肥流量传感器，利用正压气流辅助肥料流形重构，采用仿真与静态试验相结合的方式构建检测模型，模型引入料流体积作为中介变量，通过面元积分法求解肥料流量。以磷酸二铵为试验材料，通过静态试验确定肥料流等效直径与光电传感器响应电压之间的显著线性关系，基于标定试验建立了修正系数多元回归模型，最终构建了肥料流量检测模型。搭建了试验台架并开展检测方法验证试验。试验结果表明，在有气流辅助条件下，基于电压与体积反馈的颗粒肥流量检测模型（U?V?Q检测模型）平均绝对百分比误差（MAPE）不超过4.55%，均方根误差（RMSE）不大于3.82 g/s，表明基于气流辅助的颗粒肥流量光电检测法具有较高的检测精度及良好的检测稳定性。相较于无气流辅助装置U?V?Q检测模型，MAPE降低2.44%，MSD降低0.74%，此外，在有气流辅助装置情况下，U?V?Q检测模型的MAPE较基于电压与肥料流量之间转换模型（U?Q检测模型）降低9.03%。研究结果表明，单独采用气流辅助或使用U?V?Q检测模型都难以获得理想检测效果，将气流辅助技术和U?V?Q检测模型相结合，能够显著提高颗粒肥流量检测精度和稳定性，为施肥机流量实时检测提供了新思路，对精准变量施肥系统闭环控制具有重要意义。

Abstract:

During the flow of granular fertilizers, they tended to obstruct each other, and the external wheel?type fertilizer distributor further caused uneven distribution, making it challenging to accurately measure the fertilizer flow. To address this issue, an airflow?assisted photoelectric detection method that enabled more accurate measurement of granular fertilizer flow was proposed. An airflow?assisted photoelectric granular fertilizer flow sensor was designed, utilizing positive pressure airflow to assist in the reconstruction of the fertilizer flow pattern. A detection model was developed by integrating simulation and static experiments. The model incorporated fertilizer flow volume as an intermediate variable and calculated the flow rate through an area element integration algorithm. Using diammonium phosphate as the test material, static experiments revealed a significant linear relationship between the equivalent diameter of the fertilizer flow and the sensor?s response voltage. The correction coefficient?s multiple regression model was determined through calibration experiments, thus constructing the fertilizer flow detection model. To validate the effectiveness of the method, a dedicated test platform was set up for verification experiments. The test results showed that under airflow?assisted conditions, the U?V?Q detection model achieved an average absolute percentage error (MAPE) of no more than 4.55%, and a root mean square error (RMSE) of no more than 3.82 g/s. This demonstrated that the airflow?assisted photoelectric detection method for granular fertilizer flow had high detection accuracy and good stability. Compared with the U?V?Q detection model without the airflow assistance device, the MAPE was reduced by 2.44%, and the MSD was reduced by 0.74%. Furthermore, with the airflow assistance device, the MAPE of the U?V?Q detection model was 9.03% lower than that of the U?Q detection model. The research results indicated that using airflow assistance alone or the U?V?Q detection model alone did not achieve optimal detection performance. However, combining airflow assistance technology with the U?V?Q detection model significantly improved both the accuracy and stability of granular fertilizer flow detection, providing insights for real?time flow detection in fertilization machines, which was of great significance for the closed?loop control of precision variable?rate fertilization systems.