采用支持向量机算法对金刚石锯片锯切木材表面粗糙度的预测  

Prediction Model of Surface Roughness of PCD Sawing Wood with SVM

作　　者：贾娜[1] 郭佳欣 花军[1] 陈红成 Jia Na;Guo Jiaxin;Hua Jun;Chen Hongcheng(Northeast Forestry University, Harbin 150040, P. R. China;Hubei Xinyunxiang Technology Development Co., Ltd)

机构地区：[1]东北林业大学,哈尔滨150040 [2]湖北鑫运祥科技发展有限公司

出　　处：《东北林业大学学报》2019年第10期85-89,100,共6页Journal of Northeast Forestry University

基　　金：国家林业局林业科学技术推广项目(2016-34)

摘　　要：为了更准确预测木材切削加工后木材表面粗糙度,通过金刚石(PCD)锯片锯切木材试验获得不同锯切转速、进给速度、锯切厚度、木材密度时的木材表面粗糙度测量值,采用支持向量机(SVM)算法建立相应的表面粗糙度预测模型,引入网格搜索法对SVM模型参数进行优化,分析参数选取及优化对木材表面粗糙度模型精度的影响。结果表明:采用PCD锯片锯切木材时,3种影响因素对木材表面粗糙度的影响程度,由大到小依次为锯片转速、锯切厚度、进给速度,且表面粗糙度值随着锯片转速的增大而降低,随着进给速度和锯切厚度的变大而增加。参数优化后的木材表面粗糙度预测模型,更能实现木材表面粗糙度的精准预测。In order to predict wood surface roughness more accurately, wood surface roughness under different sawing speed, feed speed, sawing thickness and wood density were obtained by PCD saw blade sawing test. The support vector machine (SVM) algorithm is used to establish the corresponding surface roughness prediction model. The mesh traversal method is introduced to optimize the SVM model parameters, and the influence of parameter selection and optimization on the accuracy of wood surface roughness model is analyzed. The main and secondary influencing factors of surface roughness of three kinds of wood are saw blade speed, saw thickness and feed speed when PCD saw blade is used to cut wood, and the value of surface roughness decreases with the increase of blade speed, and increases with the increase of feed speed and sawing thickness. The prediction model of wood surface roughness after parameter optimization can achieve more accurate prediction of wood surface roughness.

关 键 词：木材表面粗糙度 预测模型 支持向量机 金刚石锯片 锯切 

分 类 号：S781.61[农业科学—木材科学与技术]