摘要
为了预测锅炉给水泵轴承温度的变化情况,提高给水泵运行的安全性和经济性,采用了统计学习理论中的核心算法——支持向量机,建立了给水泵温度预测模型(SVAR)。并通过一个实例,与基于灰色方法建立的预测模型(GM)和基于自回归方法建立的预测模型(AR)进行了比较。结果表明:基于支持向量自回归的给水泵轴承温度预测模型具有精度高、速度快、易于建模的特点。应用该方法建立的预测模型能够很好地预测给水泵运行中的温度状况,有效地避免给水泵运行中出现的故障。
In order to forecast the temperature state of water feeding pump bearings and increase the operating safety and economy of water feeding pump, the kernel algorithm of Statistical Learning Theory ( SLT), Support Vector Machine ( SVM), is applied to set up a forecast model of water feeding pump bearing temperature. The model based on SVAR is compared with the model based on Gray Model and the model based on Autoregressive by a case. The result indicate that the forecast model of water feeding pump bearing temperature based on SVAR have many advantage, such as high precision, high calculation velocity, modeling easily. The model based on SVAR can forecast the bearing temperature condition of water feeding pump in operation, and avoid the fault due to the bearing temperature variety.
出处
《轴承》
北大核心
2009年第2期53-57,共5页
Bearing
基金
水利部“948”科技创新项目(CT200516)
辽宁省教育厅科技公关项目(05L385)
关键词
滚动轴承
温度
支持向量机
统计学习理论
给水泵
预测
rolling bearing
temperature
Support Vector Machines
statistical learning theory
feed water pump
forecast