基于ε-SVR的风筛式清选装置清选性能预测研究  

Prediction for Working Performance of Air-and-screen Cleaning Unit Based on the ε-SVR Method

作　　者：梁振伟[1] 李耀明[1] 周全[1] 马征[1] 魏纯才 王建鹏 

机构地区：[1]江苏大学现代农业装备与技术教育部重点实验室,江苏镇江212013

出　　处：《农机化研究》2018年第4期26-30,36,共6页Journal of Agricultural Mechanization Research

基　　金：国家重点研发计划项目(2016YFD0702004);国家自然科学基金项目(51375214)

摘　　要：在分析传统预测模型不足之处的基础上,为了能方便地预测清选参数对清选性能的影响,将一种支持向量机模型引入风筛式清选装置的清选性能预测领域,探讨了样本容量大小对ε-SVR回归模型预测性能的影响,并与BP预测模型进行了对比。分析结果表明:采用非启发式Grid Search方法及启发式GA和PSO方法寻求ε-SVR模型最佳参数,可避免凭经验选取参数的随机性,在具有小样本的清选性能预测中,ε-SVR模型预测性能优于BP模型。On the basis of analyzing disadvantages of conventional prediction model of air-and-screen cleaning device,a new regression model based on support vector machine was proposed to predict and control of cleaning process precisely.Parameters ofε-SVR models were determined utilizing non-heuristic Grid Search、heuristic GA and PSO which could avoidthe choice of randomness.The effect of samples in different on prediction performance ofε-SVR was analyzed compared with BP.The results indicate that the prediction property ofε-SVR is better than BP especially for small sample space.

关 键 词：清选 联合收获机 风筛式 ε-SVR 

分 类 号：S225.3[农业科学—农业机械化工程]