船舶轴系支点轴承振动烈度界限值的确定和预测  

Threshold determination and prediction of ship shaft pivot bearing vibration intensity

作　　者：刘东风[1,2] 周志才[1] 孙云岭[3] 王铮[2] 

机构地区：[1]海军工程大学青岛油液检测分析中心 [2]中国人民解放军91663部队 [3]海军工程大学动力工程学院

出　　处：《上海海事大学学报》2015年第4期57-60,共4页Journal of Shanghai Maritime University

基　　金：国防装备研究基金(HW2013526;HW2014394)

摘　　要：由于船舶轴系支点轴承缺乏相应的振动烈度评判标准,结合支点轴承的日常监测数据,采用箱形图制定低转速工况下的界限值,根据振动烈度随转速增加既有线性增长又有指数增长的趋势,利用灰色线性回归组合模型对高转速工况下的界限值进行预测,并与BP神经网络模型的预测效果进行对比分析.结果表明,箱形图法制定的低转速工况界限值比较合理,灰色线性回归组合模型的预测值比BP神经网络模型的更加准确、稳健,其平均相对误差为2.634%.Due to the lack of criteria for ship shaft pivot bearing vibration intensity,Box-plot is used to draft the threshold of low rotation speed conditions with the daily pivot bearing monitoring data. According to the linear and exponential growth trends of vibration intensity with the increase of rotation speed,a grey-linear regression combined model is used to predict the threshold of high rotation speed conditions,and the prediction result is compared with that of the BP neural network model. The result shows that,the threshold of low rotation speed conditions drafted by Box-plot is reasonable,the prediction values by the grey-linear regression combined model are more accurate and robust than those by the BP neural network model,and the average relative error is 2. 634%.

关 键 词：支点轴承 振动烈度 界限值预测 箱形图 灰色线性回归组合模型 

分 类 号：U664.2[交通运输工程—船舶及航道工程] TH133.3[交通运输工程—船舶与海洋工程]