Authors

Abstract

To ensure food security and reduce harvest losses, improving the monitoring accuracy of grain combine harvester operation loss is of great importance. This paper systematically analyzes the technical progress of piezoelectric sensing applications in this field. In terms of materials, piezoelectric thin films (PVDF) exhibit faster response speeds (signal attenuation shortened by 30%), but are prone to short circuits in high-humidity environments. Piezoelectric ceramics (PZT), when combined with a double-layer vibration isolation structure, can effectively reduce vibration interference errors to below 5%, providing better stability. Regarding sensor structure, the array layout enhances multi-target recognition, while the innovative double-layer cross structure enables analytical positioning of the spatial distribution of grain collisions, offering a new approach for accurately calculating loss rates. In signal processing algorithms, support vector machines (SVM) and decision trees perform well with small sample sizes; however, combining them with discrete element simulation (EDEM) is necessary to optimize feature extraction. Among these methods, the WOA-BP algorithm can control monitoring error within 6.23% through adaptive parameter adjustment. Nevertheless, current technologies still face challenges such as insufficient adaptability to varying material environments and limited algorithm generalization under complex working conditions. In the future, multidisciplinary collaborative innovation is required to develop hybrid algorithm models that integrate weather-resistant composite materials, intelligent adaptive sensor structures, and physical mechanisms, thereby establishing a high-precision, low-cost monitoring system and providing theoretical support for the research and development of grain loss detection equipment.

Abstract in Chinese

为保障粮食安全、减少收获环节损失，提升谷物联合收获机作业损失监测精度至关重要。本文系统分析了压电效应在该领域的技术进展：材料方面，压电薄膜（PVDF）响应速度更快（信号衰减缩短30%），但高湿环境易短路；压电陶瓷（PZT）结合双层隔振结构则能有效降低振动干扰误差至5%以下，稳定性更佳。传感器结构上，阵列式布局增强多目标识别，而创新的双层十字交叉结构可实现籽粒碰撞空间分布解析定位，为损失率精准计算提供新思路。信号处理算法中，支持向量机（SVM）与决策树在小样本下表现好，但需结合离散元仿真（EDEM）优化特征提取；其中WOA-BP算法通过自适应参数调整可将监测误差控制在6.23%。然而，现有技术仍面临材料环境适应性不足及复杂工况下算法泛化能力有限等挑战。未来需多学科协同创新，开发耐候性复合材料、智能自适应传感器结构及融合物理机理的混合算法模型，以构建高精度、低成本监测系统，为粮食减损装备研发提供理论支撑。