基于蜻蜓算法的甘蔗收获机刀盘振动SVM预测模型  

Prediction Model of Cutter Vibration for Sugarcane Harvester Based on Support Vector Machine and Dragonfly Algorithm

作　　者：宋俊敏 李尚平 周永权 钟家勤[3] 

机构地区：[1]广西民族大学信息科学与工程学院,南宁530006 [2]广西高校复杂系统与智能计算重点实验室,南宁530006 [3]钦州学院机械与船舶海洋工程学院,广西钦州535000

出　　处：《农机化研究》2018年第1期20-28,共9页Journal of Agricultural Mechanization Research

基　　金：国家自然科学基金项目(51465006;61463007)

摘　　要：甘蔗收获机切割器刀盘振动是影响甘蔗宿根切割质量的一个关键因素,因此寻找复杂因素对刀盘轴向振动的影响规律并实现对刀盘振动的预测与控制有着至关重要的作用。为解决传统预测方法精度低、参数选取盲目等问题,提出一种基于蜻蜓算法的甘蔗收获机刀盘振动支持向量机预测模型。该方法利用蜻蜓群体寻优的过程实现对支持向量机参数的优化,并将优化后的支持向量机对刀盘振动进行预测。通过Mat Lab进行20次仿真实验,并与BP神经网络预测模型和传统支持向量机预测模型的预测结果进行比较,实验数据表明:基于蜻蜓算法的支持向量机预测模型具有更高的预测精度和泛化能力。结果显示:基于蜻蜓算法优化的支持向量机对刀盘振动预测的拟合率达到了99.99%,有效提高了甘蔗收获机刀盘振动的预测精度,从而表明基于蜻蜓算法优化的支持向量机预测模型对实现甘蔗收获机刀盘振动预测的有效性,为后续甘蔗收获机宿根切割质量的智能化预测及实现对甘蔗收获机减振的结构优化设计提供了有效依据。The cutter vibration of sugarcane harvester is a key factor affecting the cutting quality, so looking for the effects of properties of cutter vibration under complex factors and realize cutter vibration prediction and control plays a crucial role. In order to solve the problems that the traditional forecasting method has low precision and poor stability, a new prediction model of cutter vibration for sugarcane harvester based on support vector machine and Dragonfly algorithm is proposed. This method using the dragonfly populations find optimal process to achieve on support vector machine pa- rameter optimization, and then use the optimized support vector machine to achieve cutter vibration prediction. The MAT- LAB simulation experiment results show that, compared with the BP neural network and the traditional support vector ma- chine prediction method, the proposed support vector machine which is optimized by dragonfly algorithm has higher pre- diction precision and generalization performance. It has effectively improved the prediction precision and the prediction accuracy has reached 99.99%. So that the support vector machine prediction model based on the optimization of the dragonfly algorithm is effective to realize the prediction of the cutter vibration of sugarcane harvester. Moreover, it laid a foundation for the intelligent prediction and control of cutting quality of sugarcane harvester, and provides an effective ba- sis for the further realization of the structure design and optimization of the vibration reduction of sugarcane harvester.

关 键 词：刀盘振动 蜻蜓算法 支持向量机 参数优化 预测模型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] S225.53[自动化与计算机技术—计算机科学与技术]